Update Intel perf vendor events
Subsystems (2)
Ian Rogers
reviewercover Update Intel perf vendor events
Update Intel perf vendor events from:
https://github.com/intel/perfmon
Adds new events and updates TMA metrics to v4.8.
Add counter and experimental event tags.
Ian Rogers (37):
perf vendor events: Update alderlake events/metrics perf vendor events: Update alderlaken events/metrics perf vendor events: Add bonnell counter information perf vendor events: Update broadwell metrics add event counter
information
perf vendor events: Update broadwellde metrics add event counter
information
perf vendor events: Update broadwellx metrics add event counter
information
perf vendor events: Update cascadelakex events/metrics perf vendor events: Update elkhartlake events
perf vendor events: Add/update emeraldrapids events/metrics perf vendor events: Add goldmont counter information perf vendor events: Add goldmontplus counter information perf vendor events: Update/add grandridge events/metrics perf vendor events: Update grandridge events and add counter
information
perf vendor events: Add haswell counter information perf vendor events: Update haswellx metrics add event counter
information
perf vendor events: Add/update icelake events/metrics perf vendor events: Add/update icelakex events/metrics perf vendor events: Update ivybridge metrics add event counter
information
perf vendor events: Update ivytown metrics add event counter
information
perf vendor events: Update jaketown metrics add event counter
information
perf vendor events: Add knightslanding counter information perf vendor events: Add lunarlake counter information perf vendor events: Update meteorlake events and add counter
information
perf vendor events: Add nehalemep counter information perf vendor events: Add nehalemex counter information perf vendor events: Add/update rocketlake events/metrics perf vendor events: Update sandybridge metrics add event counter
information
perf vendor events: Add/update sapphirerapids events/metrics perf vendor events: Add/update sierraforest events/metrics perf vendor events: Add silvermont counter information perf vendor events: Add/update skylake events/metrics perf vendor events: Add/update skylakex events/metrics perf vendor events: Add snowridgex counter information perf vendor events: Add/update tigerlake events/metrics perf vendor events: Add westmereep-dp counter information perf vendor events: Add westmereep-sp counter information perf vendor events: Add westmereex counter information
.../arch/x86/alderlake/adl-metrics.json | 988 +++-- .../pmu-events/arch/x86/alderlake/cache.json | 184 +- .../arch/x86/alderlake/floating-point.json | 20 + .../arch/x86/alderlake/frontend.json | 56 +- .../pmu-events/arch/x86/alderlake/memory.json | 44 + .../arch/x86/alderlake/metricgroups.json | 23 +- .../pmu-events/arch/x86/alderlake/other.json | 37 + .../arch/x86/alderlake/pipeline.json | 214 + .../x86/alderlake/uncore-interconnect.json | 19 + .../arch/x86/alderlake/uncore-memory.json | 25 + .../arch/x86/alderlake/uncore-other.json | 1 + .../arch/x86/alderlake/virtual-memory.json | 26 + .../arch/x86/alderlaken/adln-metrics.json | 658 ++- .../pmu-events/arch/x86/alderlaken/cache.json | 50 +- .../arch/x86/alderlaken/floating-point.json | 2 + .../arch/x86/alderlaken/frontend.json | 3 + .../arch/x86/alderlaken/memory.json | 11 + .../arch/x86/alderlaken/metricgroups.json | 21 +- .../pmu-events/arch/x86/alderlaken/other.json | 13 + .../arch/x86/alderlaken/pipeline.json | 80 + .../x86/alderlaken/uncore-interconnect.json | 19 + .../arch/x86/alderlaken/uncore-memory.json | 25 + .../arch/x86/alderlaken/uncore-other.json | 1 + .../arch/x86/alderlaken/virtual-memory.json | 6 + .../pmu-events/arch/x86/bonnell/cache.json | 93 + .../pmu-events/arch/x86/bonnell/counter.json | 7 + .../arch/x86/bonnell/floating-point.json | 32 + .../pmu-events/arch/x86/bonnell/frontend.json | 11 + .../pmu-events/arch/x86/bonnell/memory.json | 19 + .../pmu-events/arch/x86/bonnell/other.json | 56 + .../pmu-events/arch/x86/bonnell/pipeline.json | 44 + .../arch/x86/bonnell/virtual-memory.json | 15 + .../arch/x86/broadwell/bdw-metrics.json | 80 +- .../pmu-events/arch/x86/broadwell/cache.json | 275 ++ .../arch/x86/broadwell/counter.json | 22 + .../arch/x86/broadwell/floating-point.json | 22 + .../arch/x86/broadwell/frontend.json | 28 + .../pmu-events/arch/x86/broadwell/memory.json | 240 ++ .../arch/x86/broadwell/metricgroups.json | 11 + .../pmu-events/arch/x86/broadwell/other.json | 4 + .../arch/x86/broadwell/pipeline.json | 137 + .../arch/x86/broadwell/uncore-cache.json | 24 + .../x86/broadwell/uncore-interconnect.json | 7 + .../arch/x86/broadwell/uncore-other.json | 10 - .../arch/x86/broadwell/virtual-memory.json | 38 + .../arch/x86/broadwellde/bdwde-metrics.json | 80 +- .../arch/x86/broadwellde/cache.json | 76 + .../arch/x86/broadwellde/counter.json | 42 + .../arch/x86/broadwellde/floating-point.json | 22 + .../arch/x86/broadwellde/frontend.json | 28 + .../arch/x86/broadwellde/memory.json | 39 + .../arch/x86/broadwellde/metricgroups.json | 11 + .../arch/x86/broadwellde/other.json | 4 + .../arch/x86/broadwellde/pipeline.json | 137 + .../arch/x86/broadwellde/uncore-cache.json | 382 ++ .../x86/broadwellde/uncore-interconnect.json | 70 + .../arch/x86/broadwellde/uncore-io.json | 62 + .../arch/x86/broadwellde/uncore-memory.json | 322 ++ .../arch/x86/broadwellde/uncore-power.json | 57 + .../arch/x86/broadwellde/virtual-memory.json | 38 + .../arch/x86/broadwellx/bdx-metrics.json | 128 +- .../pmu-events/arch/x86/broadwellx/cache.json | 88 + .../arch/x86/broadwellx/counter.json | 57 + .../arch/x86/broadwellx/floating-point.json | 22 + .../arch/x86/broadwellx/frontend.json | 28 + .../arch/x86/broadwellx/memory.json | 58 + .../arch/x86/broadwellx/metricgroups.json | 11 + .../pmu-events/arch/x86/broadwellx/other.json | 4 + .../arch/x86/broadwellx/pipeline.json | 137 + .../arch/x86/broadwellx/uncore-cache.json | 399 ++ .../x86/broadwellx/uncore-interconnect.json | 454 ++ .../arch/x86/broadwellx/uncore-io.json | 62 + .../arch/x86/broadwellx/uncore-memory.json | 326 ++ .../arch/x86/broadwellx/uncore-power.json | 57 + .../arch/x86/broadwellx/virtual-memory.json | 38 + .../arch/x86/cascadelakex/cache.json | 1245 ++++++ .../arch/x86/cascadelakex/clx-metrics.json | 310 +- .../arch/x86/cascadelakex/counter.json | 52 + .../arch/x86/cascadelakex/floating-point.json | 16 + .../arch/x86/cascadelakex/frontend.json | 49 + .../arch/x86/cascadelakex/memory.json | 743 ++++ .../arch/x86/cascadelakex/metricgroups.json | 13 + .../arch/x86/cascadelakex/other.json | 168 + .../arch/x86/cascadelakex/pipeline.json | 104 +- .../arch/x86/cascadelakex/uncore-cache.json | 2293 ++++++++++ .../x86/cascadelakex/uncore-interconnect.json | 2536 ++++++++++++ .../arch/x86/cascadelakex/uncore-io.json | 703 ++++ .../arch/x86/cascadelakex/uncore-memory.json | 985 +++++ .../arch/x86/cascadelakex/uncore-power.json | 50 + .../arch/x86/cascadelakex/virtual-memory.json | 28 + .../arch/x86/elkhartlake/cache.json | 101 + .../arch/x86/elkhartlake/counter.json | 7 + .../arch/x86/elkhartlake/floating-point.json | 3 + .../arch/x86/elkhartlake/frontend.json | 9 + .../arch/x86/elkhartlake/memory.json | 40 + .../arch/x86/elkhartlake/other.json | 61 + .../arch/x86/elkhartlake/pipeline.json | 60 + .../arch/x86/elkhartlake/virtual-memory.json | 31 + .../arch/x86/emeraldrapids/cache.json | 159 +- .../arch/x86/emeraldrapids/counter.json | 82 + .../arch/x86/emeraldrapids/emr-metrics.json | 2186 ++++++++++ .../x86/emeraldrapids/floating-point.json | 28 + .../arch/x86/emeraldrapids/frontend.json | 50 + .../arch/x86/emeraldrapids/memory.json | 50 + .../arch/x86/emeraldrapids/metricgroups.json | 137 + .../arch/x86/emeraldrapids/other.json | 43 + .../arch/x86/emeraldrapids/pipeline.json | 133 + .../arch/x86/emeraldrapids/uncore-cache.json | 1288 ++++++ .../arch/x86/emeraldrapids/uncore-cxl.json | 110 + .../emeraldrapids/uncore-interconnect.json | 1427 +++++++ .../arch/x86/emeraldrapids/uncore-io.json | 743 ++++ .../arch/x86/emeraldrapids/uncore-memory.json | 742 ++++ .../arch/x86/emeraldrapids/uncore-power.json | 49 + .../x86/emeraldrapids/virtual-memory.json | 20 + .../pmu-events/arch/x86/goldmont/cache.json | 103 + .../pmu-events/arch/x86/goldmont/counter.json | 7 + .../arch/x86/goldmont/floating-point.json | 3 + .../arch/x86/goldmont/frontend.json | 8 + .../pmu-events/arch/x86/goldmont/memory.json | 3 + .../pmu-events/arch/x86/goldmont/other.json | 5 + .../arch/x86/goldmont/pipeline.json | 40 + .../arch/x86/goldmont/virtual-memory.json | 7 + .../arch/x86/goldmontplus/cache.json | 101 + .../arch/x86/goldmontplus/counter.json | 7 + .../arch/x86/goldmontplus/floating-point.json | 3 + .../arch/x86/goldmontplus/frontend.json | 8 + .../arch/x86/goldmontplus/memory.json | 3 + .../arch/x86/goldmontplus/other.json | 5 + .../arch/x86/goldmontplus/pipeline.json | 42 + .../arch/x86/goldmontplus/virtual-memory.json | 18 + .../pmu-events/arch/x86/grandridge/cache.json | 97 +- .../arch/x86/grandridge/counter.json | 42 + .../arch/x86/grandridge/floating-point.json | 54 +- .../arch/x86/grandridge/frontend.json | 5 +- .../arch/x86/grandridge/grr-metrics.json | 849 ++++ .../arch/x86/grandridge/memory.json | 13 +- .../arch/x86/grandridge/metricgroups.json | 23 + .../pmu-events/arch/x86/grandridge/other.json | 15 +- .../arch/x86/grandridge/pipeline.json | 97 +- .../arch/x86/grandridge/uncore-cache.json | 267 ++ .../x86/grandridge/uncore-interconnect.json | 30 + .../arch/x86/grandridge/uncore-io.json | 181 + .../arch/x86/grandridge/uncore-memory.json | 66 + .../arch/x86/grandridge/uncore-power.json | 1 + .../arch/x86/grandridge/virtual-memory.json | 17 + .../arch/x86/graniterapids/cache.json | 825 ++++ .../arch/x86/graniterapids/counter.json | 77 + .../x86/graniterapids/floating-point.json | 242 ++ .../arch/x86/graniterapids/frontend.json | 469 ++- .../arch/x86/graniterapids/memory.json | 175 +- .../arch/x86/graniterapids/other.json | 150 +- .../arch/x86/graniterapids/pipeline.json | 1009 ++++- .../arch/x86/graniterapids/uncore-cache.json | 3674 +++++++++++++++++ .../arch/x86/graniterapids/uncore-cxl.json | 31 + .../graniterapids/uncore-interconnect.json | 1849 +++++++++ .../arch/x86/graniterapids/uncore-io.json | 1901 +++++++++ .../arch/x86/graniterapids/uncore-memory.json | 449 ++ .../arch/x86/graniterapids/uncore-power.json | 11 + .../x86/graniterapids/virtual-memory.json | 159 + .../pmu-events/arch/x86/haswell/cache.json | 94 + .../pmu-events/arch/x86/haswell/counter.json | 22 + .../arch/x86/haswell/floating-point.json | 10 + .../pmu-events/arch/x86/haswell/frontend.json | 29 + .../arch/x86/haswell/hsw-metrics.json | 66 +- .../pmu-events/arch/x86/haswell/memory.json | 60 + .../arch/x86/haswell/metricgroups.json | 11 + .../pmu-events/arch/x86/haswell/other.json | 4 + .../pmu-events/arch/x86/haswell/pipeline.json | 130 + .../arch/x86/haswell/uncore-cache.json | 33 + .../arch/x86/haswell/uncore-interconnect.json | 6 + .../arch/x86/haswell/uncore-other.json | 1 + .../arch/x86/haswell/virtual-memory.json | 49 + .../pmu-events/arch/x86/haswellx/cache.json | 97 + .../pmu-events/arch/x86/haswellx/counter.json | 57 + .../arch/x86/haswellx/floating-point.json | 10 + .../arch/x86/haswellx/frontend.json | 29 + .../arch/x86/haswellx/hsx-metrics.json | 114 +- .../pmu-events/arch/x86/haswellx/memory.json | 67 + .../arch/x86/haswellx/metricgroups.json | 11 + .../pmu-events/arch/x86/haswellx/other.json | 4 + .../arch/x86/haswellx/pipeline.json | 130 + .../arch/x86/haswellx/uncore-cache.json | 398 ++ .../x86/haswellx/uncore-interconnect.json | 448 ++ .../arch/x86/haswellx/uncore-io.json | 59 + .../arch/x86/haswellx/uncore-memory.json | 325 ++ .../arch/x86/haswellx/uncore-power.json | 62 + .../arch/x86/haswellx/virtual-memory.json | 49 + .../pmu-events/arch/x86/icelake/cache.json | 109 + .../pmu-events/arch/x86/icelake/counter.json | 17 + .../arch/x86/icelake/floating-point.json | 13 + .../pmu-events/arch/x86/icelake/frontend.json | 41 +- .../arch/x86/icelake/icl-metrics.json | 308 +- .../pmu-events/arch/x86/icelake/memory.json | 44 + .../arch/x86/icelake/metricgroups.json | 13 + .../pmu-events/arch/x86/icelake/other.json | 27 + .../pmu-events/arch/x86/icelake/pipeline.json | 94 + .../arch/x86/icelake/uncore-interconnect.json | 34 +- .../arch/x86/icelake/uncore-other.json | 1 + .../arch/x86/icelake/virtual-memory.json | 20 + .../pmu-events/arch/x86/icelakex/cache.json | 106 + .../pmu-events/arch/x86/icelakex/counter.json | 57 + .../arch/x86/icelakex/floating-point.json | 13 + .../arch/x86/icelakex/frontend.json | 38 + .../arch/x86/icelakex/icx-metrics.json | 340 +- .../pmu-events/arch/x86/icelakex/memory.json | 45 + .../arch/x86/icelakex/metricgroups.json | 13 + .../pmu-events/arch/x86/icelakex/other.json | 52 + .../arch/x86/icelakex/pipeline.json | 92 + .../arch/x86/icelakex/uncore-cache.json | 2149 +++++++++- .../x86/icelakex/uncore-interconnect.json | 3344 +++++++++++++++ .../arch/x86/icelakex/uncore-io.json | 1829 ++++++++ .../arch/x86/icelakex/uncore-memory.json | 338 ++ .../arch/x86/icelakex/uncore-power.json | 51 + .../arch/x86/icelakex/virtual-memory.json | 22 + .../pmu-events/arch/x86/ivybridge/cache.json | 104 + .../arch/x86/ivybridge/counter.json | 17 + .../arch/x86/ivybridge/floating-point.json | 17 + .../arch/x86/ivybridge/frontend.json | 30 + .../arch/x86/ivybridge/ivb-metrics.json | 68 +- .../pmu-events/arch/x86/ivybridge/memory.json | 19 + .../arch/x86/ivybridge/metricgroups.json | 11 + .../pmu-events/arch/x86/ivybridge/other.json | 4 + .../arch/x86/ivybridge/pipeline.json | 126 + .../arch/x86/ivybridge/uncore-cache.json | 25 + .../x86/ivybridge/uncore-interconnect.json | 9 + .../arch/x86/ivybridge/virtual-memory.json | 18 + .../pmu-events/arch/x86/ivytown/cache.json | 118 + .../pmu-events/arch/x86/ivytown/counter.json | 52 + .../arch/x86/ivytown/floating-point.json | 17 + .../pmu-events/arch/x86/ivytown/frontend.json | 30 + .../arch/x86/ivytown/ivt-metrics.json | 68 +- .../pmu-events/arch/x86/ivytown/memory.json | 41 + .../arch/x86/ivytown/metricgroups.json | 11 + .../pmu-events/arch/x86/ivytown/other.json | 4 + .../pmu-events/arch/x86/ivytown/pipeline.json | 126 + .../arch/x86/ivytown/uncore-cache.json | 349 ++ .../arch/x86/ivytown/uncore-interconnect.json | 385 ++ .../arch/x86/ivytown/uncore-io.json | 61 + .../arch/x86/ivytown/uncore-memory.json | 198 + .../arch/x86/ivytown/uncore-power.json | 74 + .../arch/x86/ivytown/virtual-memory.json | 20 + .../pmu-events/arch/x86/jaketown/cache.json | 123 + .../pmu-events/arch/x86/jaketown/counter.json | 52 + .../arch/x86/jaketown/floating-point.json | 15 + .../arch/x86/jaketown/frontend.json | 32 + .../arch/x86/jaketown/jkt-metrics.json | 24 +- .../pmu-events/arch/x86/jaketown/memory.json | 35 + .../arch/x86/jaketown/metricgroups.json | 11 + .../pmu-events/arch/x86/jaketown/other.json | 6 + .../arch/x86/jaketown/pipeline.json | 127 + .../arch/x86/jaketown/uncore-cache.json | 205 + .../x86/jaketown/uncore-interconnect.json | 207 + .../arch/x86/jaketown/uncore-io.json | 36 + .../arch/x86/jaketown/uncore-memory.json | 51 + .../arch/x86/jaketown/uncore-power.json | 39 + .../arch/x86/jaketown/virtual-memory.json | 16 + .../arch/x86/knightslanding/cache.json | 213 + .../arch/x86/knightslanding/counter.json | 37 + .../x86/knightslanding/floating-point.json | 3 + .../arch/x86/knightslanding/frontend.json | 7 + .../arch/x86/knightslanding/memory.json | 101 + .../arch/x86/knightslanding/pipeline.json | 45 + .../arch/x86/knightslanding/uncore-cache.json | 421 ++ .../arch/x86/knightslanding/uncore-io.json | 24 + .../x86/knightslanding/uncore-memory.json | 14 + .../x86/knightslanding/virtual-memory.json | 7 + .../pmu-events/arch/x86/lunarlake/cache.json | 20 + .../arch/x86/lunarlake/frontend.json | 3 + .../pmu-events/arch/x86/lunarlake/memory.json | 15 + .../pmu-events/arch/x86/lunarlake/other.json | 6 + .../arch/x86/lunarlake/pipeline.json | 36 + .../arch/x86/lunarlake/virtual-memory.json | 6 + tools/perf/pmu-events/arch/x86/mapfile.csv | 34 +- .../pmu-events/arch/x86/meteorlake/cache.json | 223 +- .../arch/x86/meteorlake/floating-point.json | 86 +- .../arch/x86/meteorlake/frontend.json | 69 +- .../arch/x86/meteorlake/memory.json | 62 +- .../pmu-events/arch/x86/meteorlake/other.json | 19 +- .../arch/x86/meteorlake/pipeline.json | 300 +- .../arch/x86/meteorlake/uncore-cache.json | 2 + .../x86/meteorlake/uncore-interconnect.json | 8 + .../arch/x86/meteorlake/uncore-memory.json | 16 + .../arch/x86/meteorlake/uncore-other.json | 1 + .../arch/x86/meteorlake/virtual-memory.json | 37 + .../pmu-events/arch/x86/nehalemep/cache.json | 320 ++ .../arch/x86/nehalemep/counter.json | 7 + .../arch/x86/nehalemep/floating-point.json | 28 + .../arch/x86/nehalemep/frontend.json | 3 + .../pmu-events/arch/x86/nehalemep/memory.json | 67 + .../pmu-events/arch/x86/nehalemep/other.json | 18 + .../arch/x86/nehalemep/pipeline.json | 109 + .../arch/x86/nehalemep/virtual-memory.json | 13 + .../pmu-events/arch/x86/nehalemex/cache.json | 315 ++ .../arch/x86/nehalemex/counter.json | 7 + .../arch/x86/nehalemex/floating-point.json | 28 + .../arch/x86/nehalemex/frontend.json | 3 + .../pmu-events/arch/x86/nehalemex/memory.json | 67 + .../pmu-events/arch/x86/nehalemex/other.json | 18 + .../arch/x86/nehalemex/pipeline.json | 109 + .../arch/x86/nehalemex/virtual-memory.json | 13 + .../pmu-events/arch/x86/rocketlake/cache.json | 109 + .../arch/x86/rocketlake/counter.json | 17 + .../arch/x86/rocketlake/floating-point.json | 13 + .../arch/x86/rocketlake/frontend.json | 41 +- .../arch/x86/rocketlake/memory.json | 44 + .../arch/x86/rocketlake/metricgroups.json | 13 + .../pmu-events/arch/x86/rocketlake/other.json | 27 + .../arch/x86/rocketlake/pipeline.json | 94 + .../arch/x86/rocketlake/rkl-metrics.json | 308 +- .../x86/rocketlake/uncore-interconnect.json | 28 +- .../arch/x86/rocketlake/uncore-other.json | 1 + .../arch/x86/rocketlake/virtual-memory.json | 20 + .../arch/x86/sandybridge/cache.json | 173 + .../arch/x86/sandybridge/counter.json | 17 + .../arch/x86/sandybridge/floating-point.json | 15 + .../arch/x86/sandybridge/frontend.json | 32 + .../arch/x86/sandybridge/memory.json | 37 + .../arch/x86/sandybridge/metricgroups.json | 11 + .../arch/x86/sandybridge/other.json | 6 + .../arch/x86/sandybridge/pipeline.json | 128 + .../arch/x86/sandybridge/snb-metrics.json | 24 +- .../arch/x86/sandybridge/uncore-cache.json | 25 + .../x86/sandybridge/uncore-interconnect.json | 9 + .../arch/x86/sandybridge/virtual-memory.json | 16 + .../arch/x86/sapphirerapids/cache.json | 161 +- .../arch/x86/sapphirerapids/counter.json | 82 + .../x86/sapphirerapids/floating-point.json | 28 + .../arch/x86/sapphirerapids/frontend.json | 50 + .../arch/x86/sapphirerapids/memory.json | 50 + .../arch/x86/sapphirerapids/metricgroups.json | 13 + .../arch/x86/sapphirerapids/other.json | 48 + .../arch/x86/sapphirerapids/pipeline.json | 133 + .../arch/x86/sapphirerapids/spr-metrics.json | 411 +- .../arch/x86/sapphirerapids/uncore-cache.json | 1244 ++++++ .../arch/x86/sapphirerapids/uncore-cxl.json | 110 + .../sapphirerapids/uncore-interconnect.json | 1427 +++++++ .../arch/x86/sapphirerapids/uncore-io.json | 679 +++ .../x86/sapphirerapids/uncore-memory.json | 742 ++++ .../arch/x86/sapphirerapids/uncore-power.json | 49 + .../x86/sapphirerapids/virtual-memory.json | 20 + .../arch/x86/sierraforest/cache.json | 97 +- .../arch/x86/sierraforest/counter.json | 77 + .../arch/x86/sierraforest/floating-point.json | 54 +- .../arch/x86/sierraforest/frontend.json | 5 +- .../arch/x86/sierraforest/memory.json | 13 +- .../arch/x86/sierraforest/metricgroups.json | 23 + .../arch/x86/sierraforest/other.json | 15 +- .../arch/x86/sierraforest/pipeline.json | 97 +- .../arch/x86/sierraforest/srf-metrics.json | 927 +++++ .../arch/x86/sierraforest/uncore-cache.json | 549 ++- .../arch/x86/sierraforest/uncore-cxl.json | 21 + .../x86/sierraforest/uncore-interconnect.json | 267 ++ .../arch/x86/sierraforest/uncore-io.json | 267 ++ .../arch/x86/sierraforest/uncore-memory.json | 66 + .../arch/x86/sierraforest/uncore-power.json | 1 + .../arch/x86/sierraforest/virtual-memory.json | 17 + .../pmu-events/arch/x86/silvermont/cache.json | 77 + .../arch/x86/silvermont/counter.json | 7 + .../arch/x86/silvermont/floating-point.json | 1 + .../arch/x86/silvermont/frontend.json | 8 + .../arch/x86/silvermont/memory.json | 1 + .../pmu-events/arch/x86/silvermont/other.json | 2 + .../arch/x86/silvermont/pipeline.json | 34 + .../arch/x86/silvermont/virtual-memory.json | 7 + .../pmu-events/arch/x86/skylake/cache.json | 250 ++ .../pmu-events/arch/x86/skylake/counter.json | 22 + .../arch/x86/skylake/floating-point.json | 10 + .../pmu-events/arch/x86/skylake/frontend.json | 49 + .../pmu-events/arch/x86/skylake/memory.json | 131 + .../arch/x86/skylake/metricgroups.json | 13 + .../pmu-events/arch/x86/skylake/other.json | 2 + .../pmu-events/arch/x86/skylake/pipeline.json | 103 +- .../arch/x86/skylake/skl-metrics.json | 196 +- .../arch/x86/skylake/uncore-cache.json | 23 + .../arch/x86/skylake/uncore-interconnect.json | 8 + .../arch/x86/skylake/uncore-other.json | 10 - .../arch/x86/skylake/virtual-memory.json | 28 + .../pmu-events/arch/x86/skylakex/cache.json | 155 + .../pmu-events/arch/x86/skylakex/counter.json | 52 + .../arch/x86/skylakex/floating-point.json | 13 + .../arch/x86/skylakex/frontend.json | 49 + .../pmu-events/arch/x86/skylakex/memory.json | 115 + .../arch/x86/skylakex/metricgroups.json | 13 + .../pmu-events/arch/x86/skylakex/other.json | 15 + .../arch/x86/skylakex/pipeline.json | 104 +- .../arch/x86/skylakex/skx-metrics.json | 310 +- .../arch/x86/skylakex/uncore-cache.json | 2274 ++++++++++ .../x86/skylakex/uncore-interconnect.json | 2521 +++++++++++ .../arch/x86/skylakex/uncore-io.json | 703 ++++ .../arch/x86/skylakex/uncore-memory.json | 804 ++++ .../arch/x86/skylakex/uncore-power.json | 50 + .../arch/x86/skylakex/virtual-memory.json | 28 + .../pmu-events/arch/x86/snowridgex/cache.json | 101 + .../arch/x86/snowridgex/counter.json | 47 + .../arch/x86/snowridgex/floating-point.json | 3 + .../arch/x86/snowridgex/frontend.json | 9 + .../arch/x86/snowridgex/memory.json | 40 + .../pmu-events/arch/x86/snowridgex/other.json | 61 + .../arch/x86/snowridgex/pipeline.json | 60 + .../arch/x86/snowridgex/uncore-cache.json | 1548 ++++++- .../x86/snowridgex/uncore-interconnect.json | 1403 +++++++ .../arch/x86/snowridgex/uncore-io.json | 1743 ++++++++ .../arch/x86/snowridgex/uncore-memory.json | 103 + .../arch/x86/snowridgex/uncore-power.json | 51 + .../arch/x86/snowridgex/virtual-memory.json | 31 + .../pmu-events/arch/x86/tigerlake/cache.json | 73 + .../arch/x86/tigerlake/counter.json | 17 + .../arch/x86/tigerlake/floating-point.json | 13 + .../arch/x86/tigerlake/frontend.json | 41 +- .../pmu-events/arch/x86/tigerlake/memory.json | 24 + .../arch/x86/tigerlake/metricgroups.json | 13 + .../pmu-events/arch/x86/tigerlake/other.json | 4 + .../arch/x86/tigerlake/pipeline.json | 95 + .../arch/x86/tigerlake/tgl-metrics.json | 198 +- .../x86/tigerlake/uncore-interconnect.json | 19 + .../arch/x86/tigerlake/uncore-memory.json | 6 + .../arch/x86/tigerlake/uncore-other.json | 1 + .../arch/x86/tigerlake/virtual-memory.json | 20 + .../arch/x86/westmereep-dp/cache.json | 282 ++ .../arch/x86/westmereep-dp/counter.json | 7 + .../x86/westmereep-dp/floating-point.json | 28 + .../arch/x86/westmereep-dp/frontend.json | 3 + .../arch/x86/westmereep-dp/memory.json | 69 + .../arch/x86/westmereep-dp/other.json | 28 + .../arch/x86/westmereep-dp/pipeline.json | 111 + .../x86/westmereep-dp/virtual-memory.json | 21 + .../arch/x86/westmereep-sp/cache.json | 321 ++ .../arch/x86/westmereep-sp/counter.json | 7 + .../x86/westmereep-sp/floating-point.json | 28 + .../arch/x86/westmereep-sp/frontend.json | 3 + .../arch/x86/westmereep-sp/memory.json | 67 + .../arch/x86/westmereep-sp/other.json | 28 + .../arch/x86/westmereep-sp/pipeline.json | 111 + .../x86/westmereep-sp/virtual-memory.json | 18 + .../pmu-events/arch/x86/westmereex/cache.json | 320 ++ .../arch/x86/westmereex/counter.json | 7 + .../arch/x86/westmereex/floating-point.json | 28 + .../arch/x86/westmereex/frontend.json | 3 + .../arch/x86/westmereex/memory.json | 68 + .../pmu-events/arch/x86/westmereex/other.json | 28 + .../arch/x86/westmereex/pipeline.json | 111 + .../arch/x86/westmereex/virtual-memory.json | 21 + 442 files changed, 76139 insertions(+), 2883 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/bonnell/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/broadwell/counter.json delete mode 100644 tools/perf/pmu-events/arch/x86/broadwell/uncore-other.json create mode 100644 tools/perf/pmu-events/arch/x86/broadwellde/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/broadwellx/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/cascadelakex/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/emeraldrapids/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/emeraldrapids/emr-metrics.json create mode 100644 tools/perf/pmu-events/arch/x86/emeraldrapids/metricgroups.json create mode 100644 tools/perf/pmu-events/arch/x86/goldmont/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/goldmontplus/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/grandridge/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/grandridge/grr-metrics.json create mode 100644 tools/perf/pmu-events/arch/x86/grandridge/metricgroups.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/floating-point.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-cache.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-cxl.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-interconnect.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-io.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.json create mode 100644 tools/perf/pmu-events/arch/x86/graniterapids/uncore-power.json create mode 100644 tools/perf/pmu-events/arch/x86/haswell/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/haswellx/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/icelake/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/icelakex/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/ivybridge/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/ivytown/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/jaketown/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/knightslanding/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/nehalemep/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/nehalemex/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/rocketlake/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/sandybridge/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/sapphirerapids/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/sierraforest/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/sierraforest/metricgroups.json create mode 100644 tools/perf/pmu-events/arch/x86/sierraforest/srf-metrics.json create mode 100644 tools/perf/pmu-events/arch/x86/silvermont/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/skylake/counter.json delete mode 100644 tools/perf/pmu-events/arch/x86/skylake/uncore-other.json create mode 100644 tools/perf/pmu-events/arch/x86/skylakex/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/snowridgex/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/tigerlake/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/westmereep-dp/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/westmereep-sp/counter.json create mode 100644 tools/perf/pmu-events/arch/x86/westmereex/counter.json
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Ian Rogers
reviewerPATCH 1/37 perf vendor events: Update alderlake events/metrics
Update events from v1.24 to v1.27.
Update p-core TMA metrics from v4.7 to v4.8, and the e-core TMA metrics to v3.6.
Bring in the event updates v1.27:
https://github.com/intel/perfmon/commit/ea4f309a04c50ca77a00da2db130fd7cf06db978
v1.26:
https://github.com/intel/perfmon/commit/0052e68d24d9873d5ff22363677794fa3eb05313
The p-core TMA 4.8 information was updated in:
https://github.com/intel/perfmon/commit/59194d4d90ca50a3fcb2de0d82b9f6fc0c9a5736
And e-core in:
https://github.com/intel/perfmon/commit/d9c2faa70bafe03129dc10f9fe414ef03a95acd9
New events are:
EXE_ACTIVITY.2_3_PORTS_UTIL, ICACHE_DATA.STALL_PERIODS,
L2_TRANS.L2_WB, MEM_TRANS_RETIRED.LOAD_LATENCY_GT_1024, MEM_UOPS_RETIRED.LOCK_LOADS, OFFCORE_REQUESTS.DEMAND_CODE_RD, OFFCORE_REQUESTS.DEMAND_RFO, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE_RD, OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD, RS.EMPTY_RESOURCE,
SERIALIZATION.C01_MS_SCB, SW_PREFETCH_ACCESS.ANY,
UOPS_ISSUED.ANY,
UOPS_ISSUED.CYCLES
--- .../arch/x86/alderlake/adl-metrics.json | 988 +++++++++--------- .../pmu-events/arch/x86/alderlake/cache.json | 184 +++- .../arch/x86/alderlake/floating-point.json | 20 + .../arch/x86/alderlake/frontend.json | 56 +- .../pmu-events/arch/x86/alderlake/memory.json | 44 + .../arch/x86/alderlake/metricgroups.json | 23 +- .../pmu-events/arch/x86/alderlake/other.json | 37 + .../arch/x86/alderlake/pipeline.json | 214 ++++ .../x86/alderlake/uncore-interconnect.json | 19 + .../arch/x86/alderlake/uncore-memory.json | 25 + .../arch/x86/alderlake/uncore-other.json | 1 + .../arch/x86/alderlake/virtual-memory.json | 26 + tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 13 files changed, 1115 insertions(+), 524 deletions(-)
tools/perf/pmu-events/arch/x86/alderlake/adl-metrics.json +472 -516
@@ -113,42 +113,30 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to certain allocation restrictions.", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.ALLOC_RESTRICTIONS@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_alloc_restriction", - "MetricThreshold": "tma_alloc_restriction > 0.1", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to certain allocation restrictions", + "MetricExpr": "tma_core_bound", + "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricName": "tma_allocation_restriction", + "MetricThreshold": "tma_allocation_restriction > 0.1 & (tma_core_bound > 0.1 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", + "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.ALL@ / tma_info_core_slots", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.ALL@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "Default;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.1", "MetricgroupNoGroup": "TopdownL1;Default", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count. The rest of these subevents count backend stalls, in cycles, due to an outstanding request which is memory bound vs core bound. The subevents are not slot based events and therefore can not be precisely added or subtracted from the Backend_Bound_Aux subevents which are slot based.", - "ScaleUnit": "100%", - "Unit": "cpu_atom" - }, - { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", - "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "tma_backend_bound", - "MetricGroup": "Default;TopdownL1;tma_L1_group", - "MetricName": "tma_backend_bound_aux", - "MetricThreshold": "tma_backend_bound_aux > 0.2", - "MetricgroupNoGroup": "TopdownL1;Default", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that UOPS must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count. All of these subevents count backend stalls, in slots, due to a resource limitation. These are not cycle based events and therefore can not be precisely added or subtracted from the Backend_Bound subevents which are cycle based. These subevents are supplementary to Backend_Bound and can be used to analyze results from a resource perspective at allocation.", + "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a mispredicted jump or a machine clear", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "(tma_info_core_slots - (cpu_atom@TOPDOWN_FE_BOUND.ALL@ + cpu_atom@TOPDOWN_BE_BOUND.ALL@ + cpu_atom@TOPDOWN_RETIRING.ALL@)) / tma_info_core_slots", + "MetricExpr": "(5 * cpu_atom@CPU_CLK_UNHALTED.CORE@ - (cpu_atom@TOPDOWN_FE_BOUND.ALL@ + cpu_atom@TOPDOWN_BE_BOUND.ALL@ + cpu_atom@TOPDOWN_RETIRING.ALL@)) / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "Default;TopdownL1;tma_L1_group", "MetricName": "tma_bad_speculation", "MetricThreshold": "tma_bad_speculation > 0.15",
@@ -158,644 +146,564 @@ "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of uops that are not from the microsequencer.", - "MetricExpr": "(cpu_atom@TOPDOWN_RETIRING.ALL@ - cpu_atom@UOPS_RETIRED.MS@) / tma_info_core_slots", - "MetricGroup": "TopdownL2;tma_L2_group;tma_retiring_group", - "MetricName": "tma_base", - "MetricThreshold": "tma_base > 0.6", - "MetricgroupNoGroup": "TopdownL2", - "ScaleUnit": "100%", - "Unit": "cpu_atom" - }, - { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.BRANCH_DETECT@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.BRANCH_DETECT@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", "MetricName": "tma_branch_detect", - "MetricThreshold": "tma_branch_detect > 0.05", - "PublicDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend. Includes BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.", + "MetricThreshold": "tma_branch_detect > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", + "PublicDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend. Includes BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to branch mispredicts.", - "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.MISPREDICT@ / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to branch mispredicts", + "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.MISPREDICT@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "TopdownL2;tma_L2_group;tma_bad_speculation_group", "MetricName": "tma_branch_mispredicts", - "MetricThreshold": "tma_branch_mispredicts > 0.05", + "MetricThreshold": "tma_branch_mispredicts > 0.05 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BTCLEARS, which occurs when the Branch Target Buffer (BTB) predicts a taken branch.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.BRANCH_RESTEER@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BTCLEARS, which occurs when the Branch Target Buffer (BTB) predicts a taken branch.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.BRANCH_RESTEER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", "MetricName": "tma_branch_resteer", - "MetricThreshold": "tma_branch_resteer > 0.05", + "MetricThreshold": "tma_branch_resteer > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to the microcode sequencer (MS).", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.CISC@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to the microcode sequencer (MS).", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.CISC@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_cisc", - "MetricThreshold": "tma_cisc > 0.05", + "MetricThreshold": "tma_cisc > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles due to backend bound stalls that are core execution bound and not attributed to outstanding demand load or store stalls.", - "MetricExpr": "max(0, tma_backend_bound - tma_memory_bound)", + "BriefDescription": "Counts the number of cycles due to backend bound stalls that are bounded by core restrictions and not attributed to an outstanding load or stores, or resource limitation", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.ALLOC_RESTRICTIONS@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", "MetricName": "tma_core_bound", - "MetricThreshold": "tma_core_bound > 0.1", + "MetricThreshold": "tma_core_bound > 0.1 & tma_backend_bound > 0.1", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to decode stalls.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.DECODE@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to decode stalls.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.DECODE@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_decode", - "MetricThreshold": "tma_decode > 0.05", + "MetricThreshold": "tma_decode > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to memory disambiguation.", - "MetricExpr": "tma_nuke * (cpu_atom@MACHINE_CLEARS.DISAMBIGUATION@ / cpu_atom@MACHINE_CLEARS.SLOW@)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_disambiguation", - "MetricThreshold": "tma_disambiguation > 0.02", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear that does not require the use of microcode, classified as a fast nuke, due to memory ordering, memory disambiguation and memory renaming", + "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.FASTNUKE@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", + "MetricName": "tma_fast_nuke", + "MetricThreshold": "tma_fast_nuke > 0.05 & (tma_machine_clears > 0.05 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load miss which hit in DRAM or MMIO (Non-DRAM).", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_DRAM_HIT@ / tma_info_core_clks - max((cpu_atom@MEM_BOUND_STALLS.LOAD@ - cpu_atom@LD_HEAD.L1_MISS_AT_RET@) / tma_info_core_clks, 0) * cpu_atom@MEM_BOUND_STALLS.LOAD_DRAM_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_dram_bound", - "MetricThreshold": "tma_dram_bound > 0.1", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to frontend stalls.", + "DefaultMetricgroupName": "TopdownL1", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ALL@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "Default;TopdownL1;tma_L1_group", + "MetricName": "tma_frontend_bound", + "MetricThreshold": "tma_frontend_bound > 0.2", + "MetricgroupNoGroup": "TopdownL1;Default", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear classified as a fast nuke due to memory ordering, memory disambiguation and memory renaming.", - "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.FASTNUKE@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", - "MetricName": "tma_fast_nuke", - "MetricThreshold": "tma_fast_nuke > 0.05", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to instruction cache misses.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ICACHE@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", + "MetricName": "tma_icache_misses", + "MetricThreshold": "tma_icache_misses > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.FRONTEND_BANDWIDTH@ / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.FRONTEND_BANDWIDTH@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "TopdownL2;tma_L2_group;tma_frontend_bound_group", - "MetricName": "tma_fetch_bandwidth", - "MetricThreshold": "tma_fetch_bandwidth > 0.1", + "MetricName": "tma_ifetch_bandwidth", + "MetricThreshold": "tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.FRONTEND_LATENCY@ / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend latency restrictions due to icache misses, itlb misses, branch detection, and resteer limitations.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.FRONTEND_LATENCY@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "TopdownL2;tma_L2_group;tma_frontend_bound_group", - "MetricName": "tma_fetch_latency", - "MetricThreshold": "tma_fetch_latency > 0.15", + "MetricName": "tma_ifetch_latency", + "MetricThreshold": "tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to FP assists.", - "MetricExpr": "tma_nuke * (cpu_atom@MACHINE_CLEARS.FP_ASSIST@ / cpu_atom@MACHINE_CLEARS.SLOW@)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_fp_assist", - "MetricThreshold": "tma_fp_assist > 0.02", - "ScaleUnit": "100%", - "Unit": "cpu_atom" - }, - { - "BriefDescription": "Counts the number of floating point divide operations per uop.", - "MetricExpr": "cpu_atom@UOPS_RETIRED.FPDIV@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_base_group", - "MetricName": "tma_fpdiv_uops", - "MetricThreshold": "tma_fpdiv_uops > 0.2", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of time that retirement is stalled due to a first level data TLB miss", + "MetricExpr": "100 * (cpu_atom@LD_HEAD.DTLB_MISS_AT_RET@ + cpu_atom@LD_HEAD.PGWALK_AT_RET@) / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricName": "tma_info_bottleneck_%_dtlb_miss_bound_cycles", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to frontend stalls.", - "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ALL@ / tma_info_core_slots", - "MetricGroup": "Default;TopdownL1;tma_L1_group", - "MetricName": "tma_frontend_bound", - "MetricThreshold": "tma_frontend_bound > 0.2", - "MetricgroupNoGroup": "TopdownL1;Default", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of time that allocation and retirement is stalled by the Frontend Cluster due to an Ifetch Miss, either Icache or ITLB Miss", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "Ifetch", + "MetricName": "tma_info_bottleneck_%_ifetch_miss_bound_cycles", + "PublicDescription": "Percentage of time that allocation and retirement is stalled by the Frontend Cluster due to an Ifetch Miss, either Icache or ITLB Miss. See Info.Ifetch_Bound", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to instruction cache misses.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ICACHE@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", - "MetricName": "tma_icache_misses", - "MetricThreshold": "tma_icache_misses > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of time that retirement is stalled due to an L1 miss", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.LOAD@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "Load_Store_Miss", + "MetricName": "tma_info_bottleneck_%_load_miss_bound_cycles", + "PublicDescription": "Percentage of time that retirement is stalled due to an L1 miss. See Info.Load_Miss_Bound", "Unit": "cpu_atom" }, { - "BriefDescription": "", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE@", - "MetricName": "tma_info_core_clks", + "BriefDescription": "Percentage of time that retirement is stalled by the Memory Cluster due to a pipeline stall", + "MetricExpr": "100 * cpu_atom@LD_HEAD.ANY_AT_RET@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "Mem_Exec", + "MetricName": "tma_info_bottleneck_%_mem_exec_bound_cycles", + "PublicDescription": "Percentage of time that retirement is stalled by the Memory Cluster due to a pipeline stall. See Info.Mem_Exec_Bound", "Unit": "cpu_atom" }, { - "BriefDescription": "", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE_P@", - "MetricName": "tma_info_core_clks_p", + "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_INST_RETIRED.ALL_BRANCHES@", + "MetricName": "tma_info_br_inst_mix_ipbranch", "Unit": "cpu_atom" }, { - "BriefDescription": "Cycles Per Instruction", - "MetricExpr": "tma_info_core_clks / INST_RETIRED.ANY", - "MetricName": "tma_info_core_cpi", + "BriefDescription": "Instruction per (near) call (lower number means higher occurrence rate)", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_INST_RETIRED.CALL@", + "MetricName": "tma_info_br_inst_mix_ipcall", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions Per Cycle", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / tma_info_core_clks", - "MetricName": "tma_info_core_ipc", + "BriefDescription": "Instructions per Far Branch ( Far Branches apply upon transition from application to operating system, handling interrupts, exceptions) [lower number means higher occurrence rate]", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_INST_RETIRED.FAR_BRANCH@u", + "MetricName": "tma_info_br_inst_mix_ipfarbranch", "Unit": "cpu_atom" }, { - "BriefDescription": "", - "MetricExpr": "5 * tma_info_core_clks", - "MetricName": "tma_info_core_slots", + "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was not taken", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / (cpu_atom@BR_MISP_RETIRED.COND@ - cpu_atom@BR_MISP_RETIRED.COND_TAKEN@)", + "MetricName": "tma_info_br_inst_mix_ipmisp_cond_ntaken", "Unit": "cpu_atom" }, { - "BriefDescription": "Uops Per Instruction", - "MetricExpr": "cpu_atom@UOPS_RETIRED.ALL@ / INST_RETIRED.ANY", - "MetricName": "tma_info_core_upi", + "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was taken", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_MISP_RETIRED.COND_TAKEN@", + "MetricName": "tma_info_br_inst_mix_ipmisp_cond_taken", "Unit": "cpu_atom" }, { - "BriefDescription": "Percent of instruction miss cost that hit in DRAM", - "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_DRAM_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", - "MetricName": "tma_info_frontend_inst_miss_cost_dramhit_percent", + "BriefDescription": "Instructions per retired indirect call or jump Branch Misprediction", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_MISP_RETIRED.INDIRECT@", + "MetricName": "tma_info_br_inst_mix_ipmisp_indirect", "Unit": "cpu_atom" }, { - "BriefDescription": "Percent of instruction miss cost that hit in the L2", - "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_L2_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", - "MetricName": "tma_info_frontend_inst_miss_cost_l2hit_percent", + "BriefDescription": "Instructions per retired return Branch Misprediction", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_MISP_RETIRED.RETURN@", + "MetricName": "tma_info_br_inst_mix_ipmisp_ret", "Unit": "cpu_atom" }, { - "BriefDescription": "Percent of instruction miss cost that hit in the L3", - "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_LLC_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", - "MetricName": "tma_info_frontend_inst_miss_cost_l3hit_percent", + "BriefDescription": "Instructions per retired Branch Misprediction", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@BR_MISP_RETIRED.ALL_BRANCHES@", + "MetricName": "tma_info_br_inst_mix_ipmispredict", "Unit": "cpu_atom" }, { "BriefDescription": "Ratio of all branches which mispredict", - "MetricExpr": "cpu_atom@BR_MISP_RETIRED.ALL_BRANCHES@ / BR_INST_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_branch_mispredict_ratio", + "MetricExpr": "cpu_atom@BR_MISP_RETIRED.ALL_BRANCHES@ / cpu_atom@BR_INST_RETIRED.ALL_BRANCHES@", + "MetricName": "tma_info_br_mispredict_bound_branch_mispredict_ratio", "Unit": "cpu_atom" }, { "BriefDescription": "Ratio between Mispredicted branches and unknown branches", - "MetricExpr": "cpu_atom@BR_MISP_RETIRED.ALL_BRANCHES@ / BACLEARS.ANY", - "MetricName": "tma_info_inst_mix_branch_mispredict_to_unknown_branch_ratio", + "MetricExpr": "cpu_atom@BR_MISP_RETIRED.ALL_BRANCHES@ / [email protected]@", + "MetricName": "tma_info_br_mispredict_bound_branch_mispredict_to_unknown_branch_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of all uops which are FPDiv uops", - "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.FPDIV@ / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_fpdiv_uop_ratio", + "BriefDescription": "Percentage of time that allocation is stalled due to load buffer full", + "MetricExpr": "100 * cpu_atom@MEM_SCHEDULER_BLOCK.LD_BUF@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricName": "tma_info_buffer_stalls_%_load_buffer_stall_cycles", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of all uops which are IDiv uops", - "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.IDIV@ / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_idiv_uop_ratio", + "BriefDescription": "Percentage of time that allocation is stalled due to memory reservation stations full", + "MetricExpr": "100 * cpu_atom@MEM_SCHEDULER_BLOCK.RSV@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricName": "tma_info_buffer_stalls_%_mem_rsv_stall_cycles", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_INST_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_ipbranch", + "BriefDescription": "Percentage of time that allocation is stalled due to store buffer full", + "MetricExpr": "100 * cpu_atom@MEM_SCHEDULER_BLOCK.ST_BUF@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricName": "tma_info_buffer_stalls_%_store_buffer_stall_cycles", "Unit": "cpu_atom" }, { - "BriefDescription": "Instruction per (near) call (lower number means higher occurrence rate)", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_INST_RETIRED.CALL", - "MetricName": "tma_info_inst_mix_ipcall", + "BriefDescription": "Cycles Per Instruction", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_core_cpi", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per Far Branch", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / (cpu_atom@BR_INST_RETIRED.FAR_BRANCH@ / 2)", - "MetricName": "tma_info_inst_mix_ipfarbranch", + "BriefDescription": "Instructions Per Cycle", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricName": "tma_info_core_ipc", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per Load", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_inst_mix_ipload", + "BriefDescription": "Uops Per Instruction", + "MetricExpr": "cpu_atom@UOPS_RETIRED.ALL@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_core_upi", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was not taken", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / (cpu_atom@BR_MISP_RETIRED.COND@ - cpu_atom@BR_MISP_RETIRED.COND_TAKEN@)", - "MetricName": "tma_info_inst_mix_ipmisp_cond_ntaken", + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss hits in the L2", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_L2_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l2hit", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was taken", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_MISP_RETIRED.COND_TAKEN", - "MetricName": "tma_info_inst_mix_ipmisp_cond_taken", + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss hits in the L3", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_LLC_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l3hit", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per retired indirect call or jump Branch Misprediction", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_MISP_RETIRED.INDIRECT", - "MetricName": "tma_info_inst_mix_ipmisp_indirect", + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss subsequently misses in the L3", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.IFETCH_DRAM_HIT@ / cpu_atom@MEM_BOUND_STALLS.IFETCH@", + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l3miss", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per retired return Branch Misprediction", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_MISP_RETIRED.RETURN", - "MetricName": "tma_info_inst_mix_ipmisp_ret", + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that hit the L2", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.LOAD_L2_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l2hit", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per retired Branch Misprediction", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / BR_MISP_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_ipmispredict", + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that hit the L3", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.LOAD_LLC_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l3hit", "Unit": "cpu_atom" }, { - "BriefDescription": "Instructions per Store", - "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / MEM_UOPS_RETIRED.ALL_STORES", - "MetricName": "tma_info_inst_mix_ipstore", + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that subsequently misses the L3", + "MetricExpr": "100 * cpu_atom@MEM_BOUND_STALLS.LOAD_DRAM_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l3miss", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of all uops which are ucode ops", - "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.MS@ / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_microcode_uop_ratio", + "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a pipeline block", + "MetricExpr": "100 * cpu_atom@LD_HEAD.L1_BOUND_AT_RET@ / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_l1_bound", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of all uops which are x87 uops", - "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.X87@ / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_x87_uop_ratio", + "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement", + "MetricExpr": "100 * (cpu_atom@LD_HEAD.L1_BOUND_AT_RET@ + cpu_atom@MEM_BOUND_STALLS.LOAD@) / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_load_bound", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of total non-speculative loads with a address aliasing block", - "MetricExpr": "100 * cpu_atom@LD_BLOCKS.4K_ALIAS@ / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_address_alias_blocks", + "BriefDescription": "Counts the number of cycles the core is stalled due to store buffer full", + "MetricExpr": "100 * (cpu_atom@MEM_SCHEDULER_BLOCK.ST_BUF@ / cpu_atom@MEM_SCHEDULER_BLOCK.ALL@) * tma_mem_scheduler", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_store_bound", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of total non-speculative loads that are splits", - "MetricExpr": "100 * cpu_atom@MEM_UOPS_RETIRED.SPLIT_LOADS@ / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_load_splits", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory disambiguation", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.DISAMBIGUATION@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_disamb_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "Percentage of total non-speculative loads with a store forward or unknown store address block", - "MetricExpr": "100 * cpu_atom@LD_BLOCKS.DATA_UNKNOWN@ / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_store_fwd_blocks", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to floating point assists", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.FP_ASSIST@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_fp_assist_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "Cycle cost per DRAM hit", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_DRAM_HIT@ / MEM_LOAD_UOPS_RETIRED.DRAM_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_dram_hit", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory ordering", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.MEMORY_ORDERING@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_monuke_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "Cycle cost per L2 hit", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_L2_HIT@ / MEM_LOAD_UOPS_RETIRED.L2_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_l2_hit", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory renaming", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.MRN_NUKE@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_mrn_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "Cycle cost per LLC hit", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_LLC_HIT@ / MEM_LOAD_UOPS_RETIRED.L3_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_l3_hit", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to page faults", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.PAGE_FAULT@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_page_fault_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "load ops retired per 1000 instruction", - "MetricExpr": "1e3 * cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@ / INST_RETIRED.ANY", - "MetricName": "tma_info_memory_memloadpki", + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to self-modifying code", + "MetricExpr": "1e3 * cpu_atom@MACHINE_CLEARS.SMC@ / cpu_atom@INST_RETIRED.ANY@", + "MetricName": "tma_info_machine_clear_bound_machine_clears_smc_pki", "Unit": "cpu_atom" }, { - "BriefDescription": "Average CPU Utilization", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", - "MetricName": "tma_info_system_cpu_utilization", + "BriefDescription": "Percentage of total non-speculative loads with an address aliasing block", + "MetricExpr": "100 * cpu_atom@LD_BLOCKS.4K_ALIAS@ / cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@", + "MetricName": "tma_info_mem_exec_blocks_%_loads_with_adressaliasing", "Unit": "cpu_atom" }, { - "BriefDescription": "Fraction of cycles spent in Kernel mode", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE@k / CPU_CLK_UNHALTED.CORE", - "MetricGroup": "Summary", - "MetricName": "tma_info_system_kernel_utilization", + "BriefDescription": "Percentage of total non-speculative loads with a store forward or unknown store address block", + "MetricExpr": "100 * cpu_atom@LD_BLOCKS.DATA_UNKNOWN@ / cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@", + "MetricName": "tma_info_mem_exec_blocks_%_loads_with_storefwdblk", "Unit": "cpu_atom" }, { - "BriefDescription": "Average Frequency Utilization relative nominal frequency", - "MetricExpr": "tma_info_core_clks / CPU_CLK_UNHALTED.REF_TSC", - "MetricGroup": "Power", - "MetricName": "tma_info_system_turbo_utilization", + "BriefDescription": "Percentage of Memory Execution Bound due to a first level data cache miss", + "MetricExpr": "100 * cpu_atom@LD_HEAD.L1_MISS_AT_RET@ / cpu_atom@LD_HEAD.ANY_AT_RET@", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_l1miss", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to Instruction Table Lookaside Buffer (ITLB) misses.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ITLB@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", - "MetricName": "tma_itlb_misses", - "MetricThreshold": "tma_itlb_misses > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of Memory Execution Bound due to other block cases, such as pipeline conflicts, fences, etc", + "MetricExpr": "100 * cpu_atom@LD_HEAD.OTHER_AT_RET@ / cpu_atom@LD_HEAD.ANY_AT_RET@", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_otherpipelineblks", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a load block.", - "MetricExpr": "cpu_atom@LD_HEAD.L1_BOUND_AT_RET@ / tma_info_core_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l1_bound", - "MetricThreshold": "tma_l1_bound > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of Memory Execution Bound due to a pagewalk", + "MetricExpr": "100 * cpu_atom@LD_HEAD.PGWALK_AT_RET@ / cpu_atom@LD_HEAD.ANY_AT_RET@", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_pagewalk", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles a core is stalled due to a demand load which hit in the L2 Cache.", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_L2_HIT@ / tma_info_core_clks - max((cpu_atom@MEM_BOUND_STALLS.LOAD@ - cpu_atom@LD_HEAD.L1_MISS_AT_RET@) / tma_info_core_clks, 0) * cpu_atom@MEM_BOUND_STALLS.LOAD_L2_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l2_bound", - "MetricThreshold": "tma_l2_bound > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of Memory Execution Bound due to a second level TLB miss", + "MetricExpr": "100 * cpu_atom@LD_HEAD.DTLB_MISS_AT_RET@ / cpu_atom@LD_HEAD.ANY_AT_RET@", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_stlbhit", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles a core is stalled due to a demand load which hit in the Last Level Cache (LLC) or other core with HITE/F/M.", - "MetricExpr": "cpu_atom@MEM_BOUND_STALLS.LOAD_LLC_HIT@ / tma_info_core_clks - max((cpu_atom@MEM_BOUND_STALLS.LOAD@ - cpu_atom@LD_HEAD.L1_MISS_AT_RET@) / tma_info_core_clks, 0) * cpu_atom@MEM_BOUND_STALLS.LOAD_LLC_HIT@ / cpu_atom@MEM_BOUND_STALLS.LOAD@", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l3_bound", - "MetricThreshold": "tma_l3_bound > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of Memory Execution Bound due to a store forward address match", + "MetricExpr": "100 * cpu_atom@LD_HEAD.ST_ADDR_AT_RET@ / cpu_atom@LD_HEAD.ANY_AT_RET@", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_storefwding", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to load buffer full", - "MetricExpr": "tma_mem_scheduler * cpu_atom@MEM_SCHEDULER_BLOCK.LD_BUF@ / MEM_SCHEDULER_BLOCK.ALL", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_ld_buffer", - "MetricThreshold": "tma_ld_buffer > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Instructions per Load", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@", + "MetricName": "tma_info_mem_mix_ipload", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a machine clear (nuke) of any kind including memory ordering and memory disambiguation.", - "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.MACHINE_CLEARS@ / tma_info_core_slots", - "MetricGroup": "TopdownL2;tma_L2_group;tma_bad_speculation_group", - "MetricName": "tma_machine_clears", - "MetricThreshold": "tma_machine_clears > 0.05", - "MetricgroupNoGroup": "TopdownL2", - "ScaleUnit": "100%", + "BriefDescription": "Instructions per Store", + "MetricExpr": "cpu_atom@INST_RETIRED.ANY@ / cpu_atom@MEM_UOPS_RETIRED.ALL_STORES@", + "MetricName": "tma_info_mem_mix_ipstore", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to memory reservation stalls in which a scheduler is not able to accept uops.", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.MEM_SCHEDULER@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_mem_scheduler", - "MetricThreshold": "tma_mem_scheduler > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of total non-speculative loads that perform one or more locks", + "MetricExpr": "100 * cpu_atom@MEM_UOPS_RETIRED.LOCK_LOADS@ / cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@", + "MetricName": "tma_info_mem_mix_load_locks_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to stores or loads.", - "MetricExpr": "min(tma_backend_bound, cpu_atom@LD_HEAD.ANY_AT_RET@ / tma_info_core_clks + tma_store_bound)", - "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", - "MetricName": "tma_memory_bound", - "MetricThreshold": "tma_memory_bound > 0.2", - "MetricgroupNoGroup": "TopdownL2", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of total non-speculative loads that are splits", + "MetricExpr": "100 * cpu_atom@MEM_UOPS_RETIRED.SPLIT_LOADS@ / cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@", + "MetricName": "tma_info_mem_mix_load_splits_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to memory ordering.", - "MetricExpr": "tma_nuke * (cpu_atom@MACHINE_CLEARS.MEMORY_ORDERING@ / cpu_atom@MACHINE_CLEARS.SLOW@)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_memory_ordering", - "MetricThreshold": "tma_memory_ordering > 0.02", - "ScaleUnit": "100%", + "BriefDescription": "Ratio of mem load uops to all uops", + "MetricExpr": "1e3 * cpu_atom@MEM_UOPS_RETIRED.ALL_LOADS@ / cpu_atom@UOPS_RETIRED.ALL@", + "MetricName": "tma_info_mem_mix_memload_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of uops that are from the complex flows issued by the micro-sequencer (MS)", - "MetricExpr": "cpu_atom@UOPS_RETIRED.MS@ / tma_info_core_slots", - "MetricGroup": "TopdownL2;tma_L2_group;tma_retiring_group", - "MetricName": "tma_ms_uops", - "MetricThreshold": "tma_ms_uops > 0.05", - "MetricgroupNoGroup": "TopdownL2", - "PublicDescription": "Counts the number of uops that are from the complex flows issued by the micro-sequencer (MS). This includes uops from flows due to complex instructions, faults, assists, and inserted flows.", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of time that the core is stalled due to a TPAUSE or UMWAIT instruction", + "MetricExpr": "100 * [email protected]_MS_SCB@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricName": "tma_info_serialization _%_tpause_cycles", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to IEC or FPC RAT stalls, which can be due to FIQ or IEC reservation stalls in which the integer, floating point or SIMD scheduler is not able to accept uops.", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.NON_MEM_SCHEDULER@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_non_mem_scheduler", - "MetricThreshold": "tma_non_mem_scheduler > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Average CPU Utilization", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear (slow nuke).", - "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.NUKE@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", - "MetricName": "tma_nuke", - "MetricThreshold": "tma_nuke > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Fraction of cycles spent in Kernel mode", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE_P@k / cpu_atom@CPU_CLK_UNHALTED.CORE@", + "MetricGroup": "Summary", + "MetricName": "tma_info_system_kernel_utilization", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to other common frontend stalls not categorized.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.OTHER@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", - "MetricName": "tma_other_fb", - "MetricThreshold": "tma_other_fb > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Average Frequency Utilization relative nominal frequency", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.CORE@ / cpu_atom@CPU_CLK_UNHALTED.REF_TSC@", + "MetricGroup": "Power", + "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a number of other load blocks.", - "MetricExpr": "cpu_atom@LD_HEAD.OTHER_AT_RET@ / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_other_l1", - "MetricThreshold": "tma_other_l1 > 0.05", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of all uops which are FPDiv uops", + "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.FPDIV@ / cpu_atom@UOPS_RETIRED.ALL@", + "MetricName": "tma_info_uop_mix_fpdiv_uop_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load miss which hits in the L2, LLC, DRAM or MMIO (Non-DRAM) but could not be correctly attributed or cycles in which the load miss is waiting on a request buffer.", - "MetricExpr": "max(0, tma_memory_bound - (tma_store_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_dram_bound))", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_other_load_store", - "MetricThreshold": "tma_other_load_store > 0.1", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of all uops which are IDiv uops", + "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.IDIV@ / cpu_atom@UOPS_RETIRED.ALL@", + "MetricName": "tma_info_uop_mix_idiv_uop_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of uops retired excluding ms and fp div uops.", - "MetricExpr": "(cpu_atom@TOPDOWN_RETIRING.ALL@ - cpu_atom@UOPS_RETIRED.MS@ - cpu_atom@UOPS_RETIRED.FPDIV@) / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_base_group", - "MetricName": "tma_other_ret", - "MetricThreshold": "tma_other_ret > 0.3", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of all uops which are microcode ops", + "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.MS@ / cpu_atom@UOPS_RETIRED.ALL@", + "MetricName": "tma_info_uop_mix_microcode_uop_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to page faults.", - "MetricExpr": "tma_nuke * (cpu_atom@MACHINE_CLEARS.PAGE_FAULT@ / cpu_atom@MACHINE_CLEARS.SLOW@)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_page_fault", - "MetricThreshold": "tma_page_fault > 0.02", - "ScaleUnit": "100%", + "BriefDescription": "Percentage of all uops which are x87 uops", + "MetricExpr": "100 * cpu_atom@UOPS_RETIRED.X87@ / cpu_atom@UOPS_RETIRED.ALL@", + "MetricName": "tma_info_uop_mix_x87_uop_ratio", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to wrong predecodes.", - "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.PREDECODE@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", - "MetricName": "tma_predecode", - "MetricThreshold": "tma_predecode > 0.05", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to Instruction Table Lookaside Buffer (ITLB) misses.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.ITLB@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", + "MetricName": "tma_itlb_misses", + "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the physical register file unable to accept an entry (marble stalls).", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.REGISTER@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_register", - "MetricThreshold": "tma_register > 0.1", + "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a machine clear (nuke) of any kind including memory ordering and memory disambiguation", + "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.MACHINE_CLEARS@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL2;tma_L2_group;tma_bad_speculation_group", + "MetricName": "tma_machine_clears", + "MetricThreshold": "tma_machine_clears > 0.05 & tma_bad_speculation > 0.15", + "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the reorder buffer being full (ROB stalls).", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.REORDER_BUFFER@ / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to memory reservation stalls in which a scheduler is not able to accept uops", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.MEM_SCHEDULER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_reorder_buffer", - "MetricThreshold": "tma_reorder_buffer > 0.1", - "ScaleUnit": "100%", - "Unit": "cpu_atom" - }, - { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", - "MetricExpr": "tma_backend_bound", - "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_aux_group", - "MetricName": "tma_resource_bound", - "MetricThreshold": "tma_resource_bound > 0.2", - "MetricgroupNoGroup": "TopdownL2", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count.", + "MetricName": "tma_mem_scheduler", + "MetricThreshold": "tma_mem_scheduler > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that result in retirement slots.", - "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_atom@TOPDOWN_RETIRING.ALL@ / tma_info_core_slots", - "MetricGroup": "Default;TopdownL1;tma_L1_group", - "MetricName": "tma_retiring", - "MetricThreshold": "tma_retiring > 0.75", - "MetricgroupNoGroup": "TopdownL1;Default", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to IEC or FPC RAT stalls, which can be due to FIQ or IEC reservation stalls in which the integer, floating point or SIMD scheduler is not able to accept uops", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.NON_MEM_SCHEDULER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", + "MetricName": "tma_non_mem_scheduler", + "MetricThreshold": "tma_non_mem_scheduler > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to RSV full relative", - "MetricExpr": "tma_mem_scheduler * cpu_atom@MEM_SCHEDULER_BLOCK.RSV@ / MEM_SCHEDULER_BLOCK.ALL", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_rsv", - "MetricThreshold": "tma_rsv > 0.05", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear that requires the use of microcode (slow nuke)", + "MetricExpr": "cpu_atom@TOPDOWN_BAD_SPECULATION.NUKE@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", + "MetricName": "tma_nuke", + "MetricThreshold": "tma_nuke > 0.05 & (tma_machine_clears > 0.05 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to scoreboards from the instruction queue (IQ), jump execution unit (JEU), or microcode sequencer (MS).", - "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.SERIALIZATION@ / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_serialization", - "MetricThreshold": "tma_serialization > 0.1", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to other common frontend stalls not categorized.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.OTHER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", + "MetricName": "tma_other_fb", + "MetricThreshold": "tma_other_fb > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to SMC.", - "MetricExpr": "tma_nuke * (cpu_atom@MACHINE_CLEARS.SMC@ / cpu_atom@MACHINE_CLEARS.SLOW@)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_smc", - "MetricThreshold": "tma_smc > 0.02", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to wrong predecodes.", + "MetricExpr": "cpu_atom@TOPDOWN_FE_BOUND.PREDECODE@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", + "MetricName": "tma_predecode", + "MetricThreshold": "tma_predecode > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to store buffer full", - "MetricExpr": "tma_store_bound", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_st_buffer", - "MetricThreshold": "tma_st_buffer > 0.05", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the physical register file unable to accept an entry (marble stalls)", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.REGISTER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", + "MetricName": "tma_register", + "MetricThreshold": "tma_register > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a first level TLB miss.", - "MetricExpr": "cpu_atom@LD_HEAD.DTLB_MISS_AT_RET@ / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_stlb_hit", - "MetricThreshold": "tma_stlb_hit > 0.05", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the reorder buffer being full (ROB stalls)", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.REORDER_BUFFER@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", + "MetricName": "tma_reorder_buffer", + "MetricThreshold": "tma_reorder_buffer > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a second level TLB miss requiring a page walk.", - "MetricExpr": "cpu_atom@LD_HEAD.PGWALK_AT_RET@ / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_stlb_miss", - "MetricThreshold": "tma_stlb_miss > 0.05", + "BriefDescription": "Counts the number of cycles the core is stalled due to a resource limitation", + "MetricExpr": "tma_backend_bound - tma_core_bound", + "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", + "MetricName": "tma_resource_bound", + "MetricThreshold": "tma_resource_bound > 0.2 & tma_backend_bound > 0.1", + "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to store buffer full.", - "MetricExpr": "tma_mem_scheduler * (cpu_atom@MEM_SCHEDULER_BLOCK.ST_BUF@ / cpu_atom@MEM_SCHEDULER_BLOCK.ALL@)", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_store_bound", - "MetricThreshold": "tma_store_bound > 0.1", + "BriefDescription": "Counts the number of issue slots that result in retirement slots", + "DefaultMetricgroupName": "TopdownL1", + "MetricExpr": "cpu_atom@TOPDOWN_RETIRING.ALL@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "Default;TopdownL1;tma_L1_group", + "MetricName": "tma_retiring", + "MetricThreshold": "tma_retiring > 0.75", + "MetricgroupNoGroup": "TopdownL1;Default", "ScaleUnit": "100%", "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a store forward block.", - "MetricExpr": "cpu_atom@LD_HEAD.ST_ADDR_AT_RET@ / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_store_fwd_blk", - "MetricThreshold": "tma_store_fwd_blk > 0.05", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to scoreboards from the instruction queue (IQ), jump execution unit (JEU), or microcode sequencer (MS)", + "MetricExpr": "cpu_atom@TOPDOWN_BE_BOUND.SERIALIZATION@ / (5 * cpu_atom@CPU_CLK_UNHALTED.CORE@)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", + "MetricName": "tma_serialization", + "MetricThreshold": "tma_serialization > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%", "Unit": "cpu_atom" },
@@ -818,7 +726,7 @@ { "BriefDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists", "MetricExpr": "78 * [email protected]@ / tma_info_thread_slots", - "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_group", + "MetricGroup": "BvIO;TopdownL4;tma_L4_group;tma_microcode_sequencer_group", "MetricName": "tma_assists", "MetricThreshold": "tma_assists > 0.1 & (tma_microcode_sequencer > 0.05 & tma_heavy_operations > 0.1)", "PublicDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists. Assists are long sequences of uops that are required in certain corner-cases for operations that cannot be handled natively by the execution pipeline. For example; when working with very small floating point values (so-called Denormals); the FP units are not set up to perform these operations natively. Instead; a sequence of instructions to perform the computation on the Denormals is injected into the pipeline. Since these microcode sequences might be dozens of uops long; Assists can be extremely deleterious to performance and they can be avoided in many cases. Sample with: ASSISTS.ANY",
@@ -838,7 +746,7 @@ "BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", - "MetricGroup": "Default;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", "MetricgroupNoGroup": "TopdownL1;Default",
@@ -861,7 +769,7 @@ { "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Branch Misprediction", "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@topdown\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", - "MetricGroup": "BadSpec;BrMispredicts;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -920,7 +828,7 @@ { "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses", "MetricExpr": "(25 * tma_info_system_core_frequency * (cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD@ * ([email protected]_DATA_RD.L3_HIT.SNOOP_HITM@ / ([email protected]_DATA_RD.L3_HIT.SNOOP_HITM@ + [email protected]_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD@))) + 24 * tma_info_system_core_frequency * cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@) * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", "MetricThreshold": "tma_contested_accesses > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses. Contested accesses occur when data written by one Logical Processor are read by another Logical Processor on a different Physical Core. Examples of contested accesses include synchronizations such as locks; true data sharing such as modified locked variables; and false sharing. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD;MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS. Related metrics: tma_data_sharing, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -941,7 +849,7 @@ { "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses", "MetricExpr": "24 * tma_info_system_core_frequency * (cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_NO_FWD@ + cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD@ * (1 - [email protected]_DATA_RD.L3_HIT.SNOOP_HITM@ / ([email protected]_DATA_RD.L3_HIT.SNOOP_HITM@ + [email protected]_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD@))) * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_info_thread_clks", - "MetricGroup": "Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_data_sharing", "MetricThreshold": "tma_data_sharing > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses. Data shared by multiple Logical Processors (even just read shared) may cause increased access latency due to cache coherency. Excessive data sharing can drastically harm multithreaded performance. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_NO_FWD. Related metrics: tma_contested_accesses, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -961,7 +869,7 @@ { "BriefDescription": "This metric represents fraction of cycles where the Divider unit was active", "MetricExpr": "[email protected]_ACTIVE@ / tma_info_thread_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricGroup": "BvCB;TopdownL3;tma_L3_group;tma_core_bound_group", "MetricName": "tma_divider", "MetricThreshold": "tma_divider > 0.2 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2)", "PublicDescription": "This metric represents fraction of cycles where the Divider unit was active. Divide and square root instructions are performed by the Divider unit and can take considerably longer latency than integer or Floating Point addition; subtraction; or multiplication. Sample with: ARITH.DIVIDER_ACTIVE",
@@ -994,14 +902,14 @@ "MetricGroup": "DSBmiss;FetchLat;TopdownL3;tma_L3_group;tma_fetch_latency_group;tma_issueFB", "MetricName": "tma_dsb_switches", "MetricThreshold": "tma_dsb_switches > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", - "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to switches from DSB to MITE pipelines. The DSB (decoded i-cache) is a Uop Cache where the front-end directly delivers Uops (micro operations) avoiding heavy x86 decoding. The DSB pipeline has shorter latency and delivered higher bandwidth than the MITE (legacy instruction decode pipeline). Switching between the two pipelines can cause penalties hence this metric measures the exposed penalty. Sample with: FRONTEND_RETIRED.DSB_MISS_PS. Related metrics: tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", + "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to switches from DSB to MITE pipelines. The DSB (decoded i-cache) is a Uop Cache where the front-end directly delivers Uops (micro operations) avoiding heavy x86 decoding. The DSB pipeline has shorter latency and delivered higher bandwidth than the MITE (legacy instruction decode pipeline). Switching between the two pipelines can cause penalties hence this metric measures the exposed penalty. Sample with: FRONTEND_RETIRED.DSB_MISS_PS. Related metrics: tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", "ScaleUnit": "100%", "Unit": "cpu_core" }, { "BriefDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses", "MetricExpr": "min(7 * cpu_core@DTLB_LOAD_MISSES.STLB_HIT\\,cmask\\=1@ + cpu_core@DTLB_LOAD_MISSES.WALK_ACTIVE@, max(cpu_core@CYCLE_ACTIVITY.CYCLES_MEM_ANY@ - cpu_core@MEMORY_ACTIVITY.CYCLES_L1D_MISS@, 0)) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", "MetricName": "tma_dtlb_load", "MetricThreshold": "tma_dtlb_load > 0.1 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses. TLBs (Translation Look-aside Buffers) are processor caches for recently used entries out of the Page Tables that are used to map virtual- to physical-addresses by the operating system. This metric approximates the potential delay of demand loads missing the first-level data TLB (assuming worst case scenario with back to back misses to different pages). This includes hitting in the second-level TLB (STLB) as well as performing a hardware page walk on an STLB miss. Sample with: MEM_INST_RETIRED.STLB_MISS_LOADS_PS. Related metrics: tma_dtlb_store, tma_info_bottleneck_memory_data_tlbs, tma_info_bottleneck_memory_synchronization",
@@ -1011,7 +919,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses", "MetricExpr": "(7 * cpu_core@DTLB_STORE_MISSES.STLB_HIT\\,cmask\\=1@ + cpu_core@DTLB_STORE_MISSES.WALK_ACTIVE@) / tma_info_core_core_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", "MetricName": "tma_dtlb_store", "MetricThreshold": "tma_dtlb_store > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses. As with ordinary data caching; focus on improving data locality and reducing working-set size to reduce DTLB overhead. Additionally; consider using profile-guided optimization (PGO) to collocate frequently-used data on the same page. Try using larger page sizes for large amounts of frequently-used data. Sample with: MEM_INST_RETIRED.STLB_MISS_STORES_PS. Related metrics: tma_dtlb_load, tma_info_bottleneck_memory_data_tlbs, tma_info_bottleneck_memory_synchronization",
@@ -1021,7 +929,7 @@ { "BriefDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing", "MetricExpr": "28 * tma_info_system_core_frequency * [email protected]_RFO.L3_HIT.SNOOP_HITM@ / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", "MetricName": "tma_false_sharing", "MetricThreshold": "tma_false_sharing > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing. False Sharing is a multithreading hiccup; where multiple Logical Processors contend on different data-elements mapped into the same cache line. Sample with: OCR.DEMAND_RFO.L3_HIT.SNOOP_HITM. Related metrics: tma_contested_accesses, tma_data_sharing, tma_machine_clears, tma_remote_cache",
@@ -1031,7 +939,7 @@ { "BriefDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed", "MetricExpr": "cpu_core@L1D_PEND_MISS.FB_FULL@ / tma_info_thread_clks", - "MetricGroup": "MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", + "MetricGroup": "BvMS;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", "PublicDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed. The higher the metric value; the deeper the memory hierarchy level the misses are satisfied from (metric values >1 are valid). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or external memory). Related metrics: tma_info_bottleneck_cache_memory_bandwidth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_latency, tma_streaming_stores",
@@ -1045,7 +953,7 @@ "MetricName": "tma_fetch_bandwidth", "MetricThreshold": "tma_fetch_bandwidth > 0.2", "MetricgroupNoGroup": "TopdownL2", - "PublicDescription": "This metric represents fraction of slots the CPU was stalled due to Frontend bandwidth issues. For example; inefficiencies at the instruction decoders; or restrictions for caching in the DSB (decoded uops cache) are categorized under Fetch Bandwidth. In such cases; the Frontend typically delivers suboptimal amount of uops to the Backend. Sample with: FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1_PS;FRONTEND_RETIRED.LATENCY_GE_1_PS;FRONTEND_RETIRED.LATENCY_GE_2_PS. Related metrics: tma_dsb_switches, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", + "PublicDescription": "This metric represents fraction of slots the CPU was stalled due to Frontend bandwidth issues. For example; inefficiencies at the instruction decoders; or restrictions for caching in the DSB (decoded uops cache) are categorized under Fetch Bandwidth. In such cases; the Frontend typically delivers suboptimal amount of uops to the Backend. Sample with: FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1_PS;FRONTEND_RETIRED.LATENCY_GE_1_PS;FRONTEND_RETIRED.LATENCY_GE_2_PS. Related metrics: tma_dsb_switches, tma_info_botlnk_l2_dsb_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", "ScaleUnit": "100%", "Unit": "cpu_core" },
@@ -1092,7 +1000,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ / (tma_retiring * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ / (tma_retiring * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -1102,7 +1010,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) vector uops fraction the CPU has retired aggregated across all vector widths", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@ / (tma_retiring * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.VECTOR@ / (tma_retiring * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -1134,7 +1042,7 @@ "BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend", "DefaultMetricgroupName": "TopdownL1", "MetricExpr": "cpu_core@topdown\\-fe\\-bound@ / (cpu_core@topdown\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@ + cpu_core@topdown\\-be\\-bound@) - cpu_core@INT_MISC.UOP_DROPPING@ / tma_info_thread_slots", - "MetricGroup": "Default;PGO;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvFB;BvIO;Default;PGO;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", "MetricgroupNoGroup": "TopdownL1;Default",
@@ -1145,7 +1053,7 @@ { "BriefDescription": "This metric represents fraction of slots where the CPU was retiring fused instructions -- where one uop can represent multiple contiguous instructions", "MetricExpr": "tma_light_operations * cpu_core@INST_RETIRED.MACRO_FUSED@ / (tma_retiring * tma_info_thread_slots)", - "MetricGroup": "Branches;Pipeline;TopdownL3;tma_L3_group;tma_light_operations_group", + "MetricGroup": "Branches;BvBO;Pipeline;TopdownL3;tma_L3_group;tma_light_operations_group", "MetricName": "tma_fused_instructions", "MetricThreshold": "tma_fused_instructions > 0.1 & tma_light_operations > 0.6", "PublicDescription": "This metric represents fraction of slots where the CPU was retiring fused instructions -- where one uop can represent multiple contiguous instructions. CMP+JCC or DEC+JCC are common examples of legacy fusions. {([MTL] Note new MOV+OP and Load+OP fusions appear under Other_Light_Ops in MTL!)}",
@@ -1166,7 +1074,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses", "MetricExpr": "cpu_core@ICACHE_DATA.STALLS@ / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_icache_misses", "MetricThreshold": "tma_icache_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses. Sample with: FRONTEND_RETIRED.L2_MISS_PS;FRONTEND_RETIRED.L1I_MISS_PS",
@@ -1183,7 +1091,7 @@ }, { "BriefDescription": "Instructions per retired mispredicts for conditional non-taken branches (lower number means higher occurrence rate).", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_MISP_RETIRED.COND_NTAKEN", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_MISP_RETIRED.COND_NTAKEN@", "MetricGroup": "Bad;BrMispredicts", "MetricName": "tma_info_bad_spec_ipmisp_cond_ntaken", "MetricThreshold": "tma_info_bad_spec_ipmisp_cond_ntaken < 200",
@@ -1191,7 +1099,7 @@ }, { "BriefDescription": "Instructions per retired mispredicts for conditional taken branches (lower number means higher occurrence rate).", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_MISP_RETIRED.COND_TAKEN", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_MISP_RETIRED.COND_TAKEN@", "MetricGroup": "Bad;BrMispredicts", "MetricName": "tma_info_bad_spec_ipmisp_cond_taken", "MetricThreshold": "tma_info_bad_spec_ipmisp_cond_taken < 200",
@@ -1199,7 +1107,7 @@ }, { "BriefDescription": "Instructions per retired mispredicts for indirect CALL or JMP branches (lower number means higher occurrence rate).", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_MISP_RETIRED.INDIRECT", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_MISP_RETIRED.INDIRECT@", "MetricGroup": "Bad;BrMispredicts", "MetricName": "tma_info_bad_spec_ipmisp_indirect", "MetricThreshold": "tma_info_bad_spec_ipmisp_indirect < 1e3",
@@ -1207,7 +1115,7 @@ }, { "BriefDescription": "Instructions per retired mispredicts for return branches (lower number means higher occurrence rate).", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_MISP_RETIRED.RET", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_MISP_RETIRED.RET@", "MetricGroup": "Bad;BrMispredicts", "MetricName": "tma_info_bad_spec_ipmisp_ret", "MetricThreshold": "tma_info_bad_spec_ipmisp_ret < 500",
@@ -1215,7 +1123,7 @@ }, { "BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear) (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_MISP_RETIRED.ALL_BRANCHES", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_MISP_RETIRED.ALL_BRANCHES@", "MetricGroup": "Bad;BadSpec;BrMispredicts", "MetricName": "tma_info_bad_spec_ipmispredict", "MetricThreshold": "tma_info_bad_spec_ipmispredict < 200",
@@ -1236,13 +1144,22 @@ "MetricThreshold": "tma_info_botlnk_l0_core_bound_likely > 0.5", "Unit": "cpu_core" }, + { + "BriefDescription": "Total pipeline cost of DSB (uop cache) hits - subset of the Instruction_Fetch_BW Bottleneck", + "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_lsd + tma_mite)))", + "MetricGroup": "DSB;FetchBW;tma_issueFB", + "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", + "MetricThreshold": "tma_info_botlnk_l2_dsb_bandwidth > 10", + "PublicDescription": "Total pipeline cost of DSB (uop cache) hits - subset of the Instruction_Fetch_BW Bottleneck. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", + "Unit": "cpu_core" + }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) misses - subset of the Instruction_Fetch_BW Bottleneck", "MetricExpr": "100 * (tma_fetch_latency * tma_dsb_switches / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_mite / (tma_dsb + tma_lsd + tma_mite))", "MetricGroup": "DSBmiss;Fed;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_misses", "MetricThreshold": "tma_info_botlnk_l2_dsb_misses > 10", - "PublicDescription": "Total pipeline cost of DSB (uop cache) misses - subset of the Instruction_Fetch_BW Bottleneck. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", + "PublicDescription": "Total pipeline cost of DSB (uop cache) misses - subset of the Instruction_Fetch_BW Bottleneck. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_bandwidth, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb, tma_lcp", "Unit": "cpu_core" }, {
@@ -1254,34 +1171,27 @@ "PublicDescription": "Total pipeline cost of Instruction Cache misses - subset of the Big_Code Bottleneck. Related metrics: ", "Unit": "cpu_core" }, - { - "BriefDescription": "Total pipeline cost of \"useful operations\" - the baseline operations not covered by Branching_Overhead nor Irregular_Overhead.", - "MetricExpr": "100 * (tma_retiring - (cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ + cpu_core@BR_INST_RETIRED.NEAR_CALL@) / tma_info_thread_slots - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", - "MetricGroup": "Ret", - "MetricName": "tma_info_bottleneck_base_non_br", - "MetricThreshold": "tma_info_bottleneck_base_non_br > 20", - "Unit": "cpu_core" - }, { "BriefDescription": "Total pipeline cost of instruction fetch related bottlenecks by large code footprint programs (i-side cache; TLB and BTB misses)", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_icache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", - "MetricGroup": "BigFootprint;Fed;Frontend;IcMiss;MemoryTLB", + "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_info_bottleneck_big_code", "MetricThreshold": "tma_info_bottleneck_big_code > 20", "Unit": "cpu_core" }, { - "BriefDescription": "Total pipeline cost of branch related instructions (used for program control-flow including function calls)", - "MetricExpr": "100 * ((cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ + cpu_core@BR_INST_RETIRED.NEAR_CALL@) / tma_info_thread_slots)", - "MetricGroup": "Ret", + "BriefDescription": "Total pipeline cost of instructions used for program control-flow - a subset of the Retiring category in TMA", + "MetricExpr": "100 * ((cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ + 2 * cpu_core@BR_INST_RETIRED.NEAR_CALL@ + cpu_core@INST_RETIRED.NOP@) / tma_info_thread_slots)", + "MetricGroup": "BvBO;Ret", "MetricName": "tma_info_bottleneck_branching_overhead", "MetricThreshold": "tma_info_bottleneck_branching_overhead > 5", + "PublicDescription": "Total pipeline cost of instructions used for program control-flow - a subset of the Retiring category in TMA. Examples include function calls; loops and alignments. (A lower bound)", "Unit": "cpu_core" }, { "BriefDescription": "Total pipeline cost of external Memory- or Cache-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", - "MetricGroup": "Mem;MemoryBW;Offcore;tma_issueBW", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_hit_latency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", + "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", "MetricName": "tma_info_bottleneck_cache_memory_bandwidth", "MetricThreshold": "tma_info_bottleneck_cache_memory_bandwidth > 20", "PublicDescription": "Total pipeline cost of external Memory- or Cache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full",
@@ -1289,8 +1199,8 @@ }, { "BriefDescription": "Total pipeline cost of external Memory- or Cache-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)))", - "MetricGroup": "Mem;MemoryLat;Offcore;tma_issueLat", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_hit_latency / (tma_dtlb_load + tma_fb_full + tma_l1_hit_latency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", + "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", "MetricName": "tma_info_bottleneck_cache_memory_latency", "MetricThreshold": "tma_info_bottleneck_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or Cache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_mem_latency",
@@ -1299,16 +1209,16 @@ { "BriefDescription": "Total pipeline cost when the execution is compute-bound - an estimation", "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializing_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "Cor;tma_issueComp", + "MetricGroup": "BvCB;Cor;tma_issueComp", "MetricName": "tma_info_bottleneck_compute_bound_est", "MetricThreshold": "tma_info_bottleneck_compute_bound_est > 20", "PublicDescription": "Total pipeline cost when the execution is compute-bound - an estimation. Covers Core Bound when High ILP as well as when long-latency execution units are busy. Related metrics: ", "Unit": "cpu_core" }, { - "BriefDescription": "Total pipeline cost of instruction fetch bandwidth related bottlenecks", + "BriefDescription": "Total pipeline cost of instruction fetch bandwidth related bottlenecks (when the front-end could not sustain operations delivery to the back-end)", "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - cpu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))) - tma_info_bottleneck_big_code", - "MetricGroup": "Fed;FetchBW;Frontend", + "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_info_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_info_bottleneck_instruction_fetch_bw > 20", "Unit": "cpu_core"
@@ -1316,7 +1226,7 @@ { "BriefDescription": "Total pipeline cost of irregular execution (e.g", "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + [email protected]\\,umask\\=1@ / tma_info_thread_clks * tma_ports_utilized_0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", - "MetricGroup": "Bad;Cor;Ret;tma_issueMS", + "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_info_bottleneck_irregular_overhead", "MetricThreshold": "tma_info_bottleneck_irregular_overhead > 10", "PublicDescription": "Total pipeline cost of irregular execution (e.g. FP-assists in HPC, Wait time with work imbalance multithreaded workloads, overhead in system services or virtualized environments). Related metrics: tma_microcode_sequencer, tma_ms_switches",
@@ -1324,8 +1234,8 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Translation related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_memory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_dtlb_load + tma_fb_full + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)))", - "MetricGroup": "Mem;MemoryTLB;Offcore;tma_issueTLB", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_memory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_dtlb_load + tma_fb_full + tma_l1_hit_latency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)))", + "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_info_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_info_bottleneck_memory_data_tlbs > 20", "PublicDescription": "Total pipeline cost of Memory Address Translation related bottlenecks (data-side TLBs). Related metrics: tma_dtlb_load, tma_dtlb_store, tma_info_bottleneck_memory_synchronization",
@@ -1334,7 +1244,7 @@ { "BriefDescription": "Total pipeline cost of Memory Synchronization related bottlenecks (data transfers and coherency updates across processors)", "MetricExpr": "100 * (tma_memory_bound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores - tma_store_latency)) + tma_machine_clears * (1 - tma_other_nukes / tma_other_nukes))", - "MetricGroup": "Mem;Offcore;tma_issueTLB", + "MetricGroup": "BvMS;Mem;Offcore;tma_issueTLB", "MetricName": "tma_info_bottleneck_memory_synchronization", "MetricThreshold": "tma_info_bottleneck_memory_synchronization > 10", "PublicDescription": "Total pipeline cost of Memory Synchronization related bottlenecks (data transfers and coherency updates across processors). Related metrics: tma_dtlb_load, tma_dtlb_store, tma_info_bottleneck_memory_data_tlbs",
@@ -1343,45 +1253,53 @@ { "BriefDescription": "Total pipeline cost of Branch Misprediction related bottlenecks", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetch_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", - "MetricGroup": "Bad;BadSpec;BrMispredicts;tma_issueBM", + "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_info_bottleneck_mispredictions", "MetricThreshold": "tma_info_bottleneck_mispredictions > 20", "PublicDescription": "Total pipeline cost of Branch Misprediction related bottlenecks. Related metrics: tma_branch_mispredicts, tma_info_bad_spec_branch_misprediction_cost, tma_mispredicts_resteers", "Unit": "cpu_core" }, { - "BriefDescription": "Total pipeline cost of remaining bottlenecks (apart from those listed in the Info.Bottlenecks metrics class)", - "MetricExpr": "100 - (tma_info_bottleneck_big_code + tma_info_bottleneck_instruction_fetch_bw + tma_info_bottleneck_mispredictions + tma_info_bottleneck_cache_memory_bandwidth + tma_info_bottleneck_cache_memory_latency + tma_info_bottleneck_memory_data_tlbs + tma_info_bottleneck_memory_synchronization + tma_info_bottleneck_compute_bound_est + tma_info_bottleneck_irregular_overhead + tma_info_bottleneck_branching_overhead + tma_info_bottleneck_base_non_br)", - "MetricGroup": "Cor;Offcore", + "BriefDescription": "Total pipeline cost of remaining bottlenecks in the back-end", + "MetricExpr": "100 - (tma_info_bottleneck_big_code + tma_info_bottleneck_instruction_fetch_bw + tma_info_bottleneck_mispredictions + tma_info_bottleneck_cache_memory_bandwidth + tma_info_bottleneck_cache_memory_latency + tma_info_bottleneck_memory_data_tlbs + tma_info_bottleneck_memory_synchronization + tma_info_bottleneck_compute_bound_est + tma_info_bottleneck_irregular_overhead + tma_info_bottleneck_branching_overhead + tma_info_bottleneck_useful_work)", + "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_info_bottleneck_other_bottlenecks", "MetricThreshold": "tma_info_bottleneck_other_bottlenecks > 20", - "PublicDescription": "Total pipeline cost of remaining bottlenecks (apart from those listed in the Info.Bottlenecks metrics class). Examples include data-dependencies (Core Bound when Low ILP) and other unlisted memory-related stalls.", + "PublicDescription": "Total pipeline cost of remaining bottlenecks in the back-end. Examples include data-dependencies (Core Bound when Low ILP) and other unlisted memory-related stalls.", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Total pipeline cost of \"useful operations\" - the portion of Retiring category not covered by Branching_Overhead nor Irregular_Overhead.", + "MetricExpr": "100 * (tma_retiring - (cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ + 2 * cpu_core@BR_INST_RETIRED.NEAR_CALL@ + cpu_core@INST_RETIRED.NOP@) / tma_info_thread_slots - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricGroup": "BvUW;Ret", + "MetricName": "tma_info_bottleneck_useful_work", + "MetricThreshold": "tma_info_bottleneck_useful_work > 20", "Unit": "cpu_core" }, { "BriefDescription": "Fraction of branches that are CALL or RET", - "MetricExpr": "(cpu_core@BR_INST_RETIRED.NEAR_CALL@ + cpu_core@BR_INST_RETIRED.NEAR_RETURN@) / BR_INST_RETIRED.ALL_BRANCHES", + "MetricExpr": "(cpu_core@BR_INST_RETIRED.NEAR_CALL@ + cpu_core@BR_INST_RETIRED.NEAR_RETURN@) / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@", "MetricGroup": "Bad;Branches", "MetricName": "tma_info_branches_callret", "Unit": "cpu_core" }, { "BriefDescription": "Fraction of branches that are non-taken conditionals", - "MetricExpr": "cpu_core@BR_INST_RETIRED.COND_NTAKEN@ / BR_INST_RETIRED.ALL_BRANCHES", + "MetricExpr": "cpu_core@BR_INST_RETIRED.COND_NTAKEN@ / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@", "MetricGroup": "Bad;Branches;CodeGen;PGO", "MetricName": "tma_info_branches_cond_nt", "Unit": "cpu_core" }, { "BriefDescription": "Fraction of branches that are taken conditionals", - "MetricExpr": "cpu_core@BR_INST_RETIRED.COND_TAKEN@ / BR_INST_RETIRED.ALL_BRANCHES", + "MetricExpr": "cpu_core@BR_INST_RETIRED.COND_TAKEN@ / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@", "MetricGroup": "Bad;Branches;CodeGen;PGO", "MetricName": "tma_info_branches_cond_tk", "Unit": "cpu_core" }, { "BriefDescription": "Fraction of branches that are unconditional (direct or indirect) jumps", - "MetricExpr": "(cpu_core@BR_INST_RETIRED.NEAR_TAKEN@ - cpu_core@BR_INST_RETIRED.COND_TAKEN@ - 2 * cpu_core@BR_INST_RETIRED.NEAR_CALL@) / BR_INST_RETIRED.ALL_BRANCHES", + "MetricExpr": "(cpu_core@BR_INST_RETIRED.NEAR_TAKEN@ - cpu_core@BR_INST_RETIRED.COND_TAKEN@ - 2 * cpu_core@BR_INST_RETIRED.NEAR_CALL@) / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@", "MetricGroup": "Bad;Branches", "MetricName": "tma_info_branches_jump", "Unit": "cpu_core"
@@ -1442,7 +1360,7 @@ "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_frontend_dsb_coverage", "MetricThreshold": "tma_info_frontend_dsb_coverage < 0.7 & tma_info_thread_ipc / 6 > 0.35", - "PublicDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache). Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_inst_mix_iptb, tma_lcp", + "PublicDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache). Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_inst_mix_iptb, tma_lcp", "Unit": "cpu_core" }, {
@@ -1468,7 +1386,7 @@ }, { "BriefDescription": "Instructions per non-speculative DSB miss (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / FRONTEND_RETIRED.ANY_DSB_MISS", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FRONTEND_RETIRED.ANY_DSB_MISS@", "MetricGroup": "DSBmiss;Fed", "MetricName": "tma_info_frontend_ipdsb_miss_ret", "MetricThreshold": "tma_info_frontend_ipdsb_miss_ret < 50",
@@ -1476,21 +1394,21 @@ }, { "BriefDescription": "Instructions per speculative Unknown Branch Misprediction (BAClear) (lower number means higher occurrence rate)", - "MetricExpr": "tma_info_inst_mix_instructions / BACLEARS.ANY", + "MetricExpr": "tma_info_inst_mix_instructions / [email protected]@", "MetricGroup": "Fed", "MetricName": "tma_info_frontend_ipunknown_branch", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache true code cacheline misses per kilo instruction", - "MetricExpr": "1e3 * cpu_core@FRONTEND_RETIRED.L2_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@FRONTEND_RETIRED.L2_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "IcMiss", "MetricName": "tma_info_frontend_l2mpki_code", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache speculative code cacheline misses per kilo instruction", - "MetricExpr": "1e3 * cpu_core@L2_RQSTS.CODE_RD_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.CODE_RD_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "IcMiss", "MetricName": "tma_info_frontend_l2mpki_code_all", "Unit": "cpu_core"
@@ -1512,7 +1430,7 @@ }, { "BriefDescription": "Branch instructions per taken branch.", - "MetricExpr": "cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ / BR_INST_RETIRED.NEAR_TAKEN", + "MetricExpr": "cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ / cpu_core@BR_INST_RETIRED.NEAR_TAKEN@", "MetricGroup": "Branches;Fed;PGO", "MetricName": "tma_info_inst_mix_bptkbranch", "Unit": "cpu_core"
@@ -1527,7 +1445,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ + cpu_core@FP_ARITH_INST_RETIRED.VECTOR@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10",
@@ -1554,7 +1472,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Double-Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST_RETIRED.SCALAR_DOUBLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_dp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_dp < 10",
@@ -1563,7 +1481,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Single-Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / FP_ARITH_INST_RETIRED.SCALAR_SINGLE", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST_RETIRED.SCALAR_SINGLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_sp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_sp < 10",
@@ -1572,7 +1490,7 @@ }, { "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_INST_RETIRED.ALL_BRANCHES", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@", "MetricGroup": "Branches;Fed;InsType", "MetricName": "tma_info_inst_mix_ipbranch", "MetricThreshold": "tma_info_inst_mix_ipbranch < 8",
@@ -1580,7 +1498,7 @@ }, { "BriefDescription": "Instructions per (near) call (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_INST_RETIRED.NEAR_CALL", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_INST_RETIRED.NEAR_CALL@", "MetricGroup": "Branches;Fed;PGO", "MetricName": "tma_info_inst_mix_ipcall", "MetricThreshold": "tma_info_inst_mix_ipcall < 200",
@@ -1596,7 +1514,7 @@ }, { "BriefDescription": "Instructions per Load (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / MEM_INST_RETIRED.ALL_LOADS", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@MEM_INST_RETIRED.ALL_LOADS@", "MetricGroup": "InsType", "MetricName": "tma_info_inst_mix_ipload", "MetricThreshold": "tma_info_inst_mix_ipload < 3",
@@ -1604,14 +1522,14 @@ }, { "BriefDescription": "Instructions per PAUSE (lower number means higher occurrence rate)", - "MetricExpr": "tma_info_inst_mix_instructions / CPU_CLK_UNHALTED.PAUSE_INST", + "MetricExpr": "tma_info_inst_mix_instructions / cpu_core@CPU_CLK_UNHALTED.PAUSE_INST@", "MetricGroup": "Flops;FpVector;InsType", "MetricName": "tma_info_inst_mix_ippause", "Unit": "cpu_core" }, { "BriefDescription": "Instructions per Store (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / MEM_INST_RETIRED.ALL_STORES", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@MEM_INST_RETIRED.ALL_STORES@", "MetricGroup": "InsType", "MetricName": "tma_info_inst_mix_ipstore", "MetricThreshold": "tma_info_inst_mix_ipstore < 8",
@@ -1626,12 +1544,12 @@ "Unit": "cpu_core" }, { - "BriefDescription": "Instruction per taken branch", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / BR_INST_RETIRED.NEAR_TAKEN", + "BriefDescription": "Instructions per taken branch", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@BR_INST_RETIRED.NEAR_TAKEN@", "MetricGroup": "Branches;Fed;FetchBW;Frontend;PGO;tma_issueFB", "MetricName": "tma_info_inst_mix_iptb", "MetricThreshold": "tma_info_inst_mix_iptb < 13", - "PublicDescription": "Instruction per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_lcp", + "PublicDescription": "Instructions per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_lcp", "Unit": "cpu_core" }, {
@@ -1664,13 +1582,13 @@ }, { "BriefDescription": "Fill Buffer (FB) hits per kilo instructions for retired demand loads (L1D misses that merge into ongoing miss-handling entries)", - "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_fb_hpki", "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L1 data cache [GB / sec]", "MetricExpr": "64 * [email protected]@ / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw",
@@ -1678,20 +1596,20 @@ }, { "BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads", - "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_l1mpki", "Unit": "cpu_core" }, { "BriefDescription": "L1 cache true misses per kilo instruction for all demand loads (including speculative)", - "MetricExpr": "1e3 * cpu_core@L2_RQSTS.ALL_DEMAND_DATA_RD@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.ALL_DEMAND_DATA_RD@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_l1mpki_load", "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L2 cache [GB / sec]", "MetricExpr": "64 * cpu_core@L2_LINES_IN.ALL@ / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l2_cache_fill_bw",
@@ -1699,48 +1617,55 @@ }, { "BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)", - "MetricExpr": "1e3 * (cpu_core@L2_RQSTS.REFERENCES@ - cpu_core@L2_RQSTS.MISS@) / INST_RETIRED.ANY", + "MetricExpr": "1e3 * (cpu_core@L2_RQSTS.REFERENCES@ - cpu_core@L2_RQSTS.MISS@) / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_l2hpki_all", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache hits per kilo instruction for all demand loads (including speculative)", - "MetricExpr": "1e3 * cpu_core@L2_RQSTS.DEMAND_DATA_RD_HIT@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.DEMAND_DATA_RD_HIT@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_l2hpki_load", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads", - "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L2_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L2_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Backend;CacheHits;Mem", "MetricName": "tma_info_memory_l2mpki", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache ([RKL+] true) misses per kilo instruction for all request types (including speculative)", - "MetricExpr": "1e3 * cpu_core@L2_RQSTS.MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem;Offcore", "MetricName": "tma_info_memory_l2mpki_all", "Unit": "cpu_core" }, { "BriefDescription": "L2 cache ([RKL+] true) misses per kilo instruction for all demand loads (including speculative)", - "MetricExpr": "1e3 * cpu_core@L2_RQSTS.DEMAND_DATA_RD_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.DEMAND_DATA_RD_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "CacheHits;Mem", "MetricName": "tma_info_memory_l2mpki_load", "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Offcore requests (L2 cache miss) per kilo instruction for demand RFOs", + "MetricExpr": "1e3 * cpu_core@L2_RQSTS.RFO_MISS@ / cpu_core@INST_RETIRED.ANY@", + "MetricGroup": "CacheMisses;Offcore", + "MetricName": "tma_info_memory_l2mpki_rfo", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Average per-thread data access bandwidth to the L3 cache [GB / sec]", "MetricExpr": "64 * cpu_core@OFFCORE_REQUESTS.ALL_REQUESTS@ / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW;Offcore", "MetricName": "tma_info_memory_l3_cache_access_bw", "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L3 cache [GB / sec]", "MetricExpr": "64 * cpu_core@LONGEST_LAT_CACHE.MISS@ / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l3_cache_fill_bw",
@@ -1748,21 +1673,21 @@ }, { "BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads", - "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L3_MISS@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L3_MISS@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Mem", "MetricName": "tma_info_memory_l3mpki", "Unit": "cpu_core" }, { "BriefDescription": "Average Parallel L2 cache miss data reads", - "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.DATA_RD@ / OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", + "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.DATA_RD@ / cpu_core@OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD@", "MetricGroup": "Memory_BW;Offcore", "MetricName": "tma_info_memory_latency_data_l2_mlp", "Unit": "cpu_core" }, { "BriefDescription": "Average Latency for L2 cache miss demand Loads", - "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD@ / OFFCORE_REQUESTS.DEMAND_DATA_RD", + "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD@ / cpu_core@OFFCORE_REQUESTS.DEMAND_DATA_RD@", "MetricGroup": "Memory_Lat;Offcore", "MetricName": "tma_info_memory_latency_load_l2_miss_latency", "Unit": "cpu_core"
@@ -1776,35 +1701,35 @@ }, { "BriefDescription": "Average Latency for L3 cache miss demand Loads", - "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD@ / OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", + "MetricExpr": "cpu_core@OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD@ / cpu_core@OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD@", "MetricGroup": "Memory_Lat;Offcore", "MetricName": "tma_info_memory_latency_load_l3_miss_latency", "Unit": "cpu_core" }, { "BriefDescription": "Actual Average Latency for L1 data-cache miss demand load operations (in core cycles)", - "MetricExpr": "cpu_core@L1D_PEND_MISS.PENDING@ / MEM_LOAD_COMPLETED.L1_MISS_ANY", + "MetricExpr": "cpu_core@L1D_PEND_MISS.PENDING@ / cpu_core@MEM_LOAD_COMPLETED.L1_MISS_ANY@", "MetricGroup": "Mem;MemoryBound;MemoryLat", "MetricName": "tma_info_memory_load_miss_real_latency", "Unit": "cpu_core" }, { "BriefDescription": "\"Bus lock\" per kilo instruction", - "MetricExpr": "1e3 * cpu_core@SQ_MISC.BUS_LOCK@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@SQ_MISC.BUS_LOCK@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Mem", "MetricName": "tma_info_memory_mix_bus_lock_pki", "Unit": "cpu_core" }, { "BriefDescription": "Un-cacheable retired load per kilo instruction", - "MetricExpr": "1e3 * cpu_core@MEM_LOAD_MISC_RETIRED.UC@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@MEM_LOAD_MISC_RETIRED.UC@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Mem", "MetricName": "tma_info_memory_mix_uc_load_pki", "Unit": "cpu_core" }, { "BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss", - "MetricExpr": "cpu_core@L1D_PEND_MISS.PENDING@ / L1D_PEND_MISS.PENDING_CYCLES", + "MetricExpr": "cpu_core@L1D_PEND_MISS.PENDING@ / cpu_core@L1D_PEND_MISS.PENDING_CYCLES@", "MetricGroup": "Mem;MemoryBW;MemoryBound", "MetricName": "tma_info_memory_mlp", "PublicDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-Logical Processor)",
@@ -1812,14 +1737,14 @@ }, { "BriefDescription": "STLB (2nd level TLB) code speculative misses per kilo instruction (misses of any page-size that complete the page walk)", - "MetricExpr": "1e3 * cpu_core@ITLB_MISSES.WALK_COMPLETED@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@ITLB_MISSES.WALK_COMPLETED@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Fed;MemoryTLB", "MetricName": "tma_info_memory_tlb_code_stlb_mpki", "Unit": "cpu_core" }, { "BriefDescription": "STLB (2nd level TLB) data load speculative misses per kilo instruction (misses of any page-size that complete the page walk)", - "MetricExpr": "1e3 * cpu_core@DTLB_LOAD_MISSES.WALK_COMPLETED@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@DTLB_LOAD_MISSES.WALK_COMPLETED@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Mem;MemoryTLB", "MetricName": "tma_info_memory_tlb_load_stlb_mpki", "Unit": "cpu_core"
@@ -1834,21 +1759,42 @@ }, { "BriefDescription": "STLB (2nd level TLB) data store speculative misses per kilo instruction (misses of any page-size that complete the page walk)", - "MetricExpr": "1e3 * cpu_core@DTLB_STORE_MISSES.WALK_COMPLETED@ / INST_RETIRED.ANY", + "MetricExpr": "1e3 * cpu_core@DTLB_STORE_MISSES.WALK_COMPLETED@ / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Mem;MemoryTLB", "MetricName": "tma_info_memory_tlb_store_stlb_mpki", "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is execution) per core", "MetricExpr": "cpu_core@UOPS_EXECUTED.THREAD@ / (cpu_core@UOPS_EXECUTED.CORE_CYCLES_GE_1@ / 2 if #SMT_on else cpu_core@UOPS_EXECUTED.THREAD\\,cmask\\=1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute", "Unit": "cpu_core" }, + { + "BriefDescription": "Average number of uops fetched from DSB per cycle", + "MetricExpr": "[email protected]_UOPS@ / [email protected]_CYCLES_ANY@", + "MetricGroup": "Fed;FetchBW", + "MetricName": "tma_info_pipeline_fetch_dsb", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Average number of uops fetched from LSD per cycle", + "MetricExpr": "[email protected]@ / [email protected]_ACTIVE@", + "MetricGroup": "Fed;FetchBW", + "MetricName": "tma_info_pipeline_fetch_lsd", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Average number of uops fetched from MITE per cycle", + "MetricExpr": "[email protected]_UOPS@ / [email protected]_CYCLES_ANY@", + "MetricGroup": "Fed;FetchBW", + "MetricName": "tma_info_pipeline_fetch_mite", + "Unit": "cpu_core" + }, { "BriefDescription": "Instructions per a microcode Assist invocation", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / ASSISTS.ANY", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / [email protected]@", "MetricGroup": "MicroSeq;Pipeline;Ret;Retire", "MetricName": "tma_info_pipeline_ipassist", "MetricThreshold": "tma_info_pipeline_ipassist < 100e3",
@@ -1887,14 +1833,14 @@ }, { "BriefDescription": "Average CPU Utilization (percentage)", - "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "tma_info_system_cpus_utilized / #num_cpus_online", "MetricGroup": "HPC;Summary", "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_core" }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "#num_cpus_online * tma_info_system_cpu_utilization", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized", "Unit": "cpu_core"
@@ -1925,14 +1871,14 @@ }, { "BriefDescription": "Cycles Per Instruction for the Operating System (OS) Kernel mode", - "MetricExpr": "CPU_CLK_UNHALTED.THREAD_P:k / cpu_core@INST_RETIRED.ANY_P@k", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD_P@k / cpu_core@INST_RETIRED.ANY_P@k", "MetricGroup": "OS", "MetricName": "tma_info_system_kernel_cpi", "Unit": "cpu_core" }, { "BriefDescription": "Fraction of cycles spent in the Operating System (OS) Kernel mode", - "MetricExpr": "CPU_CLK_UNHALTED.THREAD_P:k / CPU_CLK_UNHALTED.THREAD", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD_P@k / cpu_core@CPU_CLK_UNHALTED.THREAD@", "MetricGroup": "OS", "MetricName": "tma_info_system_kernel_utilization", "MetricThreshold": "tma_info_system_kernel_utilization > 0.05",
@@ -1971,7 +1917,7 @@ }, { "BriefDescription": "Average Frequency Utilization relative nominal frequency", - "MetricExpr": "tma_info_thread_clks / CPU_CLK_UNHALTED.REF_TSC", + "MetricExpr": "tma_info_thread_clks / cpu_core@CPU_CLK_UNHALTED.REF_TSC@", "MetricGroup": "Power", "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_core"
@@ -1992,7 +1938,7 @@ }, { "BriefDescription": "The ratio of Executed- by Issued-Uops", - "MetricExpr": "cpu_core@UOPS_EXECUTED.THREAD@ / UOPS_ISSUED.ANY", + "MetricExpr": "cpu_core@UOPS_EXECUTED.THREAD@ / cpu_core@UOPS_ISSUED.ANY@", "MetricGroup": "Cor;Pipeline", "MetricName": "tma_info_thread_execute_per_issue", "PublicDescription": "The ratio of Executed- by Issued-Uops. Ratio > 1 suggests high rate of uop micro-fusions. Ratio < 1 suggest high rate of \"execute\" at rename stage.",
@@ -2021,15 +1967,15 @@ }, { "BriefDescription": "Uops Per Instruction", - "MetricExpr": "tma_retiring * tma_info_thread_slots / INST_RETIRED.ANY", + "MetricExpr": "tma_retiring * tma_info_thread_slots / cpu_core@INST_RETIRED.ANY@", "MetricGroup": "Pipeline;Ret;Retire", "MetricName": "tma_info_thread_uoppi", "MetricThreshold": "tma_info_thread_uoppi > 1.05", "Unit": "cpu_core" }, { - "BriefDescription": "Instruction per taken branch", - "MetricExpr": "tma_retiring * tma_info_thread_slots / BR_INST_RETIRED.NEAR_TAKEN", + "BriefDescription": "Uops per taken branch", + "MetricExpr": "tma_retiring * tma_info_thread_slots / cpu_core@BR_INST_RETIRED.NEAR_TAKEN@", "MetricGroup": "Branches;Fed;FetchBW", "MetricName": "tma_info_thread_uptb", "MetricThreshold": "tma_info_thread_uptb < 9",
@@ -2068,7 +2014,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses", "MetricExpr": "cpu_core@ICACHE_TAG.STALLS@ / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_itlb_misses", "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses. Sample with: FRONTEND_RETIRED.STLB_MISS_PS;FRONTEND_RETIRED.ITLB_MISS_PS",
@@ -2085,10 +2031,20 @@ "ScaleUnit": "100%", "Unit": "cpu_core" }, + { + "BriefDescription": "This metric roughly estimates fraction of cycles with demand load accesses that hit the L1 cache", + "MetricExpr": "min(2 * (cpu_core@MEM_INST_RETIRED.ALL_LOADS@ - cpu_core@MEM_LOAD_RETIRED.FB_HIT@ - cpu_core@MEM_LOAD_RETIRED.L1_MISS@) * 20 / 100, max(cpu_core@CYCLE_ACTIVITY.CYCLES_MEM_ANY@ - cpu_core@MEMORY_ACTIVITY.CYCLES_L1D_MISS@, 0)) / tma_info_thread_clks", + "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound_group", + "MetricName": "tma_l1_hit_latency", + "MetricThreshold": "tma_l1_hit_latency > 0.1 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", + "PublicDescription": "This metric roughly estimates fraction of cycles with demand load accesses that hit the L1 cache. The short latency of the L1 data cache may be exposed in pointer-chasing memory access patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "ScaleUnit": "100%", + "Unit": "cpu_core" + }, { "BriefDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads", "MetricExpr": "(cpu_core@MEMORY_ACTIVITY.STALLS_L1D_MISS@ - cpu_core@MEMORY_ACTIVITY.STALLS_L2_MISS@) / tma_info_thread_clks", - "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", + "MetricGroup": "BvML;CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", "MetricName": "tma_l2_bound", "MetricThreshold": "tma_l2_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2)", "PublicDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads. Avoiding cache misses (i.e. L1 misses/L2 hits) can improve the latency and increase performance. Sample with: MEM_LOAD_RETIRED.L2_HIT_PS",
@@ -2108,7 +2064,7 @@ { "BriefDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited)", "MetricExpr": "9 * tma_info_system_core_frequency * (cpu_core@MEM_LOAD_RETIRED.L3_HIT@ * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2)) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", + "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3 hits) will improve the latency; reduce contention with sibling physical cores and increase performance. Note the value of this node may overlap with its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tma_info_bottleneck_cache_memory_latency, tma_mem_latency",
@@ -2121,7 +2077,7 @@ "MetricGroup": "FetchLat;TopdownL3;tma_L3_group;tma_fetch_latency_group;tma_issueFB", "MetricName": "tma_lcp", "MetricThreshold": "tma_lcp > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", - "PublicDescription": "This metric represents fraction of cycles CPU was stalled due to Length Changing Prefixes (LCPs). Using proper compiler flags or Intel Compiler by default will certainly avoid this. #Link: Optimization Guide about LCP BKMs. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb", + "PublicDescription": "This metric represents fraction of cycles CPU was stalled due to Length Changing Prefixes (LCPs). Using proper compiler flags or Intel Compiler by default will certainly avoid this. #Link: Optimization Guide about LCP BKMs. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_botlnk_l2_dsb_bandwidth, tma_info_botlnk_l2_dsb_misses, tma_info_frontend_dsb_coverage, tma_info_inst_mix_iptb", "ScaleUnit": "100%", "Unit": "cpu_core" },
@@ -2170,7 +2126,7 @@ "MetricGroup": "Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_l1_bound_group", "MetricName": "tma_lock_latency", "MetricThreshold": "tma_lock_latency > 0.2 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric represents fraction of cycles the CPU spent handling cache misses due to lock operations. Due to the microarchitecture handling of locks; they are classified as L1_Bound regardless of what memory source satisfied them. Sample with: MEM_INST_RETIRED.LOCK_LOADS_PS. Related metrics: tma_store_latency", + "PublicDescription": "This metric represents fraction of cycles the CPU spent handling cache misses due to lock operations. Due to the microarchitecture handling of locks; they are classified as L1_Bound regardless of what memory source satisfied them. Sample with: MEM_INST_RETIRED.LOCK_LOADS. Related metrics: tma_store_latency", "ScaleUnit": "100%", "Unit": "cpu_core" },
@@ -2187,7 +2143,7 @@ { "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Machine Clears", "MetricExpr": "max(0, tma_bad_speculation - tma_branch_mispredicts)", - "MetricGroup": "BadSpec;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", + "MetricGroup": "BadSpec;BvMS;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", "MetricName": "tma_machine_clears", "MetricThreshold": "tma_machine_clears > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -2198,7 +2154,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(cpu_core@CPU_CLK_UNHALTED.THREAD@, cpu_core@OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD\\,cmask\\=4@) / tma_info_thread_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuristic assumes that a similar off-core traffic is generated by all IA cores. This metric does not aggregate non-data-read requests by this logical processor; requests from other IA Logical Processors/Physical Cores/sockets; or other non-IA devices like GPU; hence the maximum external memory bandwidth limits may or may not be approached when this metric is flagged (see Uncore counters for that). Related metrics: tma_fb_full, tma_info_bottleneck_cache_memory_bandwidth, tma_info_system_dram_bw_use, tma_sq_full",
@@ -2208,7 +2164,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(cpu_core@CPU_CLK_UNHALTED.THREAD@, cpu_core@OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD@) / tma_info_thread_clks - tma_mem_bandwidth", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from other Logical Processors/Physical Cores/sockets (see Uncore counters for that). Related metrics: tma_info_bottleneck_cache_memory_latency, tma_l3_hit_latency",
@@ -2258,7 +2214,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage", "MetricExpr": "tma_branch_mispredicts / tma_bad_speculation * cpu_core@INT_MISC.CLEAR_RESTEER_CYCLES@ / tma_info_thread_clks", - "MetricGroup": "BadSpec;BrMispredicts;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", "MetricThreshold": "tma_mispredicts_resteers > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage. Sample with: INT_MISC.CLEAR_RESTEER_CYCLES. Related metrics: tma_branch_mispredicts, tma_info_bad_spec_branch_misprediction_cost, tma_info_bottleneck_mispredictions",
@@ -2298,7 +2254,7 @@ { "BriefDescription": "This metric represents fraction of slots where the CPU was retiring branch instructions that were not fused", "MetricExpr": "tma_light_operations * (cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ - cpu_core@INST_RETIRED.MACRO_FUSED@) / (tma_retiring * tma_info_thread_slots)", - "MetricGroup": "Branches;Pipeline;TopdownL3;tma_L3_group;tma_light_operations_group", + "MetricGroup": "Branches;BvBO;Pipeline;TopdownL3;tma_L3_group;tma_light_operations_group", "MetricName": "tma_non_fused_branches", "MetricThreshold": "tma_non_fused_branches > 0.1 & tma_light_operations > 0.6", "PublicDescription": "This metric represents fraction of slots where the CPU was retiring branch instructions that were not fused. Non-conditional branches like direct JMP or CALL would count here. Can be used to examine fusible conditional jumps that were not fused.",
@@ -2308,7 +2264,7 @@ { "BriefDescription": "This metric represents fraction of slots where the CPU was retiring NOP (no op) instructions", "MetricExpr": "tma_light_operations * cpu_core@INST_RETIRED.NOP@ / (tma_retiring * tma_info_thread_slots)", - "MetricGroup": "Pipeline;TopdownL4;tma_L4_group;tma_other_light_ops_group", + "MetricGroup": "BvBO;Pipeline;TopdownL4;tma_L4_group;tma_other_light_ops_group", "MetricName": "tma_nop_instructions", "MetricThreshold": "tma_nop_instructions > 0.1 & (tma_other_light_ops > 0.3 & tma_light_operations > 0.6)", "PublicDescription": "This metric represents fraction of slots where the CPU was retiring NOP (no op) instructions. Compilers often use NOPs for certain address alignments - e.g. start address of a function or loop body. Sample with: INST_RETIRED.NOP",
@@ -2328,7 +2284,7 @@ { "BriefDescription": "This metric estimates fraction of slots the CPU was stalled due to other cases of misprediction (non-retired x86 branches or other types).", "MetricExpr": "max(tma_branch_mispredicts * (1 - cpu_core@BR_MISP_RETIRED.ALL_BRANCHES@ / (cpu_core@INT_MISC.CLEARS_COUNT@ - cpu_core@MACHINE_CLEARS.COUNT@)), 0.0001)", - "MetricGroup": "BrMispredicts;TopdownL3;tma_L3_group;tma_branch_mispredicts_group", + "MetricGroup": "BrMispredicts;BvIO;TopdownL3;tma_L3_group;tma_branch_mispredicts_group", "MetricName": "tma_other_mispredicts", "MetricThreshold": "tma_other_mispredicts > 0.05 & (tma_branch_mispredicts > 0.1 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%",
@@ -2337,7 +2293,7 @@ { "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Nukes (Machine Clears) not related to memory ordering.", "MetricExpr": "max(tma_machine_clears * (1 - cpu_core@MACHINE_CLEARS.MEMORY_ORDERING@ / cpu_core@MACHINE_CLEARS.COUNT@), 0.0001)", - "MetricGroup": "Machine_Clears;TopdownL3;tma_L3_group;tma_machine_clears_group", + "MetricGroup": "BvIO;Machine_Clears;TopdownL3;tma_L3_group;tma_machine_clears_group", "MetricName": "tma_other_nukes", "MetricThreshold": "tma_other_nukes > 0.05 & (tma_machine_clears > 0.1 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%",
@@ -2395,7 +2351,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU executed no uops on any execution port (Logical Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "(cpu_core@EXE_ACTIVITY.3_PORTS_UTIL\\,umask\\=0x80@ + [email protected]\\,umask\\=1@) / tma_info_thread_clks * (cpu_core@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@) / tma_info_thread_clks", + "MetricExpr": "(cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + max([email protected]\\,umask\\=1@ - cpu_core@RESOURCE_STALLS.SCOREBOARD@, 0)) / tma_info_thread_clks * (cpu_core@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utilization > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))",
@@ -2428,7 +2384,7 @@ "BriefDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise)", "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@UOPS_EXECUTED.CYCLES_GE_3@ / tma_info_thread_clks", - "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", + "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", "MetricName": "tma_ports_utilized_3m", "MetricThreshold": "tma_ports_utilized_3m > 0.4 & (tma_ports_utilization > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", "PublicDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise). Sample with: UOPS_EXECUTED.CYCLES_GE_3",
@@ -2439,7 +2395,7 @@ "BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", - "MetricGroup": "Default;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.1", "MetricgroupNoGroup": "TopdownL1;Default",
@@ -2450,7 +2406,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU issue-pipeline was stalled due to serializing operations", "MetricExpr": "cpu_core@RESOURCE_STALLS.SCOREBOARD@ / tma_info_thread_clks + tma_c02_wait", - "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_group;tma_issueSO", + "MetricGroup": "BvIO;PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_group;tma_issueSO", "MetricName": "tma_serializing_operation", "MetricThreshold": "tma_serializing_operation > 0.1 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2)", "PublicDescription": "This metric represents fraction of cycles the CPU issue-pipeline was stalled due to serializing operations. Instructions like CPUID; WRMSR or LFENCE serialize the out-of-order execution which may limit performance. Sample with: RESOURCE_STALLS.SCOREBOARD. Related metrics: tma_ms_switches",
@@ -2501,7 +2457,7 @@ { "BriefDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors)", "MetricExpr": "([email protected]_CYCLES@ + cpu_core@L1D_PEND_MISS.L2_STALLS@) / tma_info_thread_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors). Related metrics: tma_fb_full, tma_info_bottleneck_cache_memory_bandwidth, tma_info_system_dram_bw_use, tma_mem_bandwidth",
@@ -2531,7 +2487,7 @@ { "BriefDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses", "MetricExpr": "(cpu_core@MEM_STORE_RETIRED.L2_HIT@ * 10 * (1 - cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ / cpu_core@MEM_INST_RETIRED.ALL_STORES@) + (1 - cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ / cpu_core@MEM_INST_RETIRED.ALL_STORES@) * min(cpu_core@CPU_CLK_UNHALTED.THREAD@, cpu_core@OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO@)) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", "MetricName": "tma_store_latency", "MetricThreshold": "tma_store_latency > 0.1 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses. Store accesses usually less impact out-of-order core performance; however; holding resources for longer time can lead into undesired implications (e.g. contention on L1D fill-buffer entries - see FB_Full). Related metrics: tma_fb_full, tma_lock_latency",
@@ -2579,7 +2535,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears", "MetricExpr": "cpu_core@INT_MISC.UNKNOWN_BRANCH_CYCLES@ / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", "MetricName": "tma_unknown_branches", "MetricThreshold": "tma_unknown_branches > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears. These are fetched branches the Branch Prediction Unit was unable to recognize (e.g. first time the branch is fetched or hitting BTB capacity limit) hence called Unknown Branches. Sample with: FRONTEND_RETIRED.UNKNOWN_BRANCH",
@@ -2588,7 +2544,7 @@ }, { "BriefDescription": "This metric serves as an approximation of legacy x87 usage", - "MetricExpr": "tma_retiring * cpu_core@UOPS_EXECUTED.X87@ / UOPS_EXECUTED.THREAD", + "MetricExpr": "tma_retiring * cpu_core@UOPS_EXECUTED.X87@ / cpu_core@UOPS_EXECUTED.THREAD@", "MetricGroup": "Compute;TopdownL4;tma_L4_group;tma_fp_arith_group", "MetricName": "tma_x87_use", "MetricThreshold": "tma_x87_use > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
tools/perf/pmu-events/arch/x86/alderlake/cache.json +182 -2
856 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlake/floating-point.json +20 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "ARITH.FPDIV_ACTIVE", + "Counter": "0,1,2,3,4,5,6,7", "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE",
@@ -10,6 +11,7 @@ }, { "BriefDescription": "Counts all microcode FP assists.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", "PublicDescription": "Counts all microcode Floating Point assists.",
@@ -19,6 +21,7 @@ }, { "BriefDescription": "ASSISTS.SSE_AVX_MIX", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", "SampleAfterValue": "1000003",
@@ -27,6 +30,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is alias to FP_ARITH_DISPATCHED.V0]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", "SampleAfterValue": "2000003",
@@ -35,6 +39,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is alias to FP_ARITH_DISPATCHED.V1]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", "SampleAfterValue": "2000003",
@@ -43,6 +48,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is alias to FP_ARITH_DISPATCHED.V2]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", "SampleAfterValue": "2000003",
@@ -51,6 +57,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias to FP_ARITH_DISPATCHED.PORT_0]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", "SampleAfterValue": "2000003",
@@ -59,6 +66,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias to FP_ARITH_DISPATCHED.PORT_1]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", "SampleAfterValue": "2000003",
@@ -67,6 +75,7 @@ }, { "BriefDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias to FP_ARITH_DISPATCHED.PORT_5]", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", "SampleAfterValue": "2000003",
@@ -75,6 +84,7 @@ }, { "BriefDescription": "Counts number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -84,6 +94,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -93,6 +104,7 @@ }, { "BriefDescription": "Counts number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -102,6 +114,7 @@ }, { "BriefDescription": "Counts number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -111,6 +124,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single and 256-bit packed double precision FP instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, 1 for each element. Applies to SSE* and AVX* packed single precision and packed double precision FP instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision and 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -120,6 +134,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP14 RSQRT14 RANGE SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", "PublicDescription": "Number of SSE/AVX computational scalar single precision and double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -129,6 +144,7 @@ }, { "BriefDescription": "Counts number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", "PublicDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -138,6 +154,7 @@ }, { "BriefDescription": "Counts number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", "PublicDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -147,6 +164,7 @@ }, { "BriefDescription": "Number of any Vector retired FP arithmetic instructions", + "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", "PublicDescription": "Number of any Vector retired FP arithmetic instructions. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -156,6 +174,7 @@ }, { "BriefDescription": "Counts the number of floating point operations retired that required microcode assist.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.FP_ASSIST", "PublicDescription": "Counts the number of floating point operations retired that required microcode assist, which is not a reflection of the number of FP operations, instructions or uops.",
@@ -165,6 +184,7 @@ }, { "BriefDescription": "Counts the number of floating point divide uops retired (x87 and SSE, including x87 sqrt).", + "Counter": "0,1,2,3,4,5", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.FPDIV", "PEBS": "1",
tools/perf/pmu-events/arch/x86/alderlake/frontend.json +55 -1
65 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlake/memory.json +44 -0
44 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlake/metricgroups.json +19 -4
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/alderlake/other.json +37 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/alderlake/pipeline.json +214 -0
214 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlake/uncore-interconnect.json +19 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json +25 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/alderlake/uncore-other.json +1 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/alderlake/virtual-memory.json +26 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
Lines of this file are not shown. View the raw message.
The rest of this diff is not shown. View the raw message.
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Part of the diff is not shown here. View the raw message.
Ian Rogers
reviewerPATCH 2/37 perf vendor events: Update alderlaken events/metrics
In reply to Ian RogersUpdate events from v1.24 to v1.27.
Update e-core TMA metrics to v3.6.
Bring in the event updates v1.27:
https://github.com/intel/perfmon/commit/ea4f309a04c50ca77a00da2db130fd7cf06db978
v1.26:
https://github.com/intel/perfmon/commit/0052e68d24d9873d5ff22363677794fa3eb05313
The e-core TMA 3.6 information was updated in:
https://github.com/intel/perfmon/commit/d9c2faa70bafe03129dc10f9fe414ef03a95acd9
New events are:
MEM_UOPS_RETIRED.LOCK_LOADS,
SERIALIZATION.C01_MS_SCB,
UOPS_ISSUED.ANY.
--- .../arch/x86/alderlaken/adln-metrics.json | 658 ++++++++---------- .../pmu-events/arch/x86/alderlaken/cache.json | 50 +- .../arch/x86/alderlaken/floating-point.json | 2 + .../arch/x86/alderlaken/frontend.json | 3 + .../arch/x86/alderlaken/memory.json | 11 + .../arch/x86/alderlaken/metricgroups.json | 21 +- .../pmu-events/arch/x86/alderlaken/other.json | 13 + .../arch/x86/alderlaken/pipeline.json | 80 +++ .../x86/alderlaken/uncore-interconnect.json | 19 + .../arch/x86/alderlaken/uncore-memory.json | 25 + .../arch/x86/alderlaken/uncore-other.json | 1 + .../arch/x86/alderlaken/virtual-memory.json | 6 + tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 13 files changed, 503 insertions(+), 388 deletions(-)
tools/perf/pmu-events/arch/x86/alderlaken/adln-metrics.json +285 -373
@@ -85,39 +85,28 @@ "ScaleUnit": "1SMI#" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to certain allocation restrictions.", - "MetricExpr": "TOPDOWN_BE_BOUND.ALLOC_RESTRICTIONS / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", - "MetricName": "tma_alloc_restriction", - "MetricThreshold": "tma_alloc_restriction > 0.1", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to certain allocation restrictions", + "MetricExpr": "tma_core_bound", + "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricName": "tma_allocation_restriction", + "MetricThreshold": "tma_allocation_restriction > 0.1 & (tma_core_bound > 0.1 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", + "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "TOPDOWN_BE_BOUND.ALL / tma_info_core_slots", + "MetricExpr": "TOPDOWN_BE_BOUND.ALL / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "Default;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.1", "MetricgroupNoGroup": "TopdownL1;Default", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count. The rest of these subevents count backend stalls, in cycles, due to an outstanding request which is memory bound vs core bound. The subevents are not slot based events and therefore can not be precisely added or subtracted from the Backend_Bound_Aux subevents which are slot based.", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", - "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "tma_backend_bound", - "MetricGroup": "Default;TopdownL1;tma_L1_group", - "MetricName": "tma_backend_bound_aux", - "MetricThreshold": "tma_backend_bound_aux > 0.2", - "MetricgroupNoGroup": "TopdownL1;Default", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that UOPS must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count. All of these subevents count backend stalls, in slots, due to a resource limitation. These are not cycle based events and therefore can not be precisely added or subtracted from the Backend_Bound subevents which are cycle based. These subevents are supplementary to Backend_Bound and can be used to analyze results from a resource perspective at allocation.", + "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count", "ScaleUnit": "100%" }, { "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a mispredicted jump or a machine clear", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "(tma_info_core_slots - (TOPDOWN_FE_BOUND.ALL + TOPDOWN_BE_BOUND.ALL + TOPDOWN_RETIRING.ALL)) / tma_info_core_slots", + "MetricExpr": "(5 * CPU_CLK_UNHALTED.CORE - (TOPDOWN_FE_BOUND.ALL + TOPDOWN_BE_BOUND.ALL + TOPDOWN_RETIRING.ALL)) / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "Default;TopdownL1;tma_L1_group", "MetricName": "tma_bad_speculation", "MetricThreshold": "tma_bad_speculation > 0.15",
@@ -126,300 +115,351 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of uops that are not from the microsequencer.", - "MetricExpr": "(TOPDOWN_RETIRING.ALL - UOPS_RETIRED.MS) / tma_info_core_slots", - "MetricGroup": "TopdownL2;tma_L2_group;tma_retiring_group", - "MetricName": "tma_base", - "MetricThreshold": "tma_base > 0.6", - "MetricgroupNoGroup": "TopdownL2", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend", - "MetricExpr": "TOPDOWN_FE_BOUND.BRANCH_DETECT / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend", + "MetricExpr": "TOPDOWN_FE_BOUND.BRANCH_DETECT / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", "MetricName": "tma_branch_detect", - "MetricThreshold": "tma_branch_detect > 0.05", - "PublicDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend. Includes BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.", + "MetricThreshold": "tma_branch_detect > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", + "PublicDescription": "Counts the number of issue slots that were not delivered by the frontend due to BACLEARS, which occurs when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend. Includes BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to branch mispredicts.", - "MetricExpr": "TOPDOWN_BAD_SPECULATION.MISPREDICT / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to branch mispredicts", + "MetricExpr": "TOPDOWN_BAD_SPECULATION.MISPREDICT / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL2;tma_L2_group;tma_bad_speculation_group", "MetricName": "tma_branch_mispredicts", - "MetricThreshold": "tma_branch_mispredicts > 0.05", + "MetricThreshold": "tma_branch_mispredicts > 0.05 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BTCLEARS, which occurs when the Branch Target Buffer (BTB) predicts a taken branch.", - "MetricExpr": "TOPDOWN_FE_BOUND.BRANCH_RESTEER / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to BTCLEARS, which occurs when the Branch Target Buffer (BTB) predicts a taken branch.", + "MetricExpr": "TOPDOWN_FE_BOUND.BRANCH_RESTEER / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", "MetricName": "tma_branch_resteer", - "MetricThreshold": "tma_branch_resteer > 0.05", + "MetricThreshold": "tma_branch_resteer > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to the microcode sequencer (MS).", - "MetricExpr": "TOPDOWN_FE_BOUND.CISC / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to the microcode sequencer (MS).", + "MetricExpr": "TOPDOWN_FE_BOUND.CISC / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_cisc", - "MetricThreshold": "tma_cisc > 0.05", + "MetricThreshold": "tma_cisc > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of cycles due to backend bound stalls that are core execution bound and not attributed to outstanding demand load or store stalls.", - "MetricExpr": "max(0, tma_backend_bound - tma_memory_bound)", + "BriefDescription": "Counts the number of cycles due to backend bound stalls that are bounded by core restrictions and not attributed to an outstanding load or stores, or resource limitation", + "MetricExpr": "TOPDOWN_BE_BOUND.ALLOC_RESTRICTIONS / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", "MetricName": "tma_core_bound", - "MetricThreshold": "tma_core_bound > 0.1", + "MetricThreshold": "tma_core_bound > 0.1 & tma_backend_bound > 0.1", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to decode stalls.", - "MetricExpr": "TOPDOWN_FE_BOUND.DECODE / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to decode stalls.", + "MetricExpr": "TOPDOWN_FE_BOUND.DECODE / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_decode", - "MetricThreshold": "tma_decode > 0.05", + "MetricThreshold": "tma_decode > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to memory disambiguation.", - "MetricExpr": "tma_nuke * (MACHINE_CLEARS.DISAMBIGUATION / MACHINE_CLEARS.SLOW)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_disambiguation", - "MetricThreshold": "tma_disambiguation > 0.02", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear that does not require the use of microcode, classified as a fast nuke, due to memory ordering, memory disambiguation and memory renaming", + "MetricExpr": "TOPDOWN_BAD_SPECULATION.FASTNUKE / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", + "MetricName": "tma_fast_nuke", + "MetricThreshold": "tma_fast_nuke > 0.05 & (tma_machine_clears > 0.05 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load miss which hit in DRAM or MMIO (Non-DRAM).", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_DRAM_HIT / tma_info_core_clks - max((MEM_BOUND_STALLS.LOAD - LD_HEAD.L1_MISS_AT_RET) / tma_info_core_clks, 0) * MEM_BOUND_STALLS.LOAD_DRAM_HIT / MEM_BOUND_STALLS.LOAD", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_dram_bound", - "MetricThreshold": "tma_dram_bound > 0.1", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to frontend stalls.", + "DefaultMetricgroupName": "TopdownL1", + "MetricExpr": "TOPDOWN_FE_BOUND.ALL / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "Default;TopdownL1;tma_L1_group", + "MetricName": "tma_frontend_bound", + "MetricThreshold": "tma_frontend_bound > 0.2", + "MetricgroupNoGroup": "TopdownL1;Default", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear classified as a fast nuke due to memory ordering, memory disambiguation and memory renaming.", - "MetricExpr": "TOPDOWN_BAD_SPECULATION.FASTNUKE / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", - "MetricName": "tma_fast_nuke", - "MetricThreshold": "tma_fast_nuke > 0.05", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to instruction cache misses.", + "MetricExpr": "TOPDOWN_FE_BOUND.ICACHE / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", + "MetricName": "tma_icache_misses", + "MetricThreshold": "tma_icache_misses > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", - "MetricExpr": "TOPDOWN_FE_BOUND.FRONTEND_BANDWIDTH / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", + "MetricExpr": "TOPDOWN_FE_BOUND.FRONTEND_BANDWIDTH / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL2;tma_L2_group;tma_frontend_bound_group", - "MetricName": "tma_fetch_bandwidth", - "MetricThreshold": "tma_fetch_bandwidth > 0.1", + "MetricName": "tma_ifetch_bandwidth", + "MetricThreshold": "tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend bandwidth restrictions due to decode, predecode, cisc, and other limitations.", - "MetricExpr": "TOPDOWN_FE_BOUND.FRONTEND_LATENCY / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to frontend latency restrictions due to icache misses, itlb misses, branch detection, and resteer limitations.", + "MetricExpr": "TOPDOWN_FE_BOUND.FRONTEND_LATENCY / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL2;tma_L2_group;tma_frontend_bound_group", - "MetricName": "tma_fetch_latency", - "MetricThreshold": "tma_fetch_latency > 0.15", + "MetricName": "tma_ifetch_latency", + "MetricThreshold": "tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to FP assists.", - "MetricExpr": "tma_nuke * (MACHINE_CLEARS.FP_ASSIST / MACHINE_CLEARS.SLOW)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_fp_assist", - "MetricThreshold": "tma_fp_assist > 0.02", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of time that retirement is stalled due to a first level data TLB miss", + "MetricExpr": "100 * (LD_HEAD.DTLB_MISS_AT_RET + LD_HEAD.PGWALK_AT_RET) / CPU_CLK_UNHALTED.CORE", + "MetricName": "tma_info_bottleneck_%_dtlb_miss_bound_cycles" }, { - "BriefDescription": "Counts the number of floating point divide operations per uop.", - "MetricExpr": "UOPS_RETIRED.FPDIV / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_base_group", - "MetricName": "tma_fpdiv_uops", - "MetricThreshold": "tma_fpdiv_uops > 0.2", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of time that allocation and retirement is stalled by the Frontend Cluster due to an Ifetch Miss, either Icache or ITLB Miss", + "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH / CPU_CLK_UNHALTED.CORE", + "MetricGroup": "Ifetch", + "MetricName": "tma_info_bottleneck_%_ifetch_miss_bound_cycles", + "PublicDescription": "Percentage of time that allocation and retirement is stalled by the Frontend Cluster due to an Ifetch Miss, either Icache or ITLB Miss. See Info.Ifetch_Bound" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to frontend stalls.", - "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "TOPDOWN_FE_BOUND.ALL / tma_info_core_slots", - "MetricGroup": "Default;TopdownL1;tma_L1_group", - "MetricName": "tma_frontend_bound", - "MetricThreshold": "tma_frontend_bound > 0.2", - "MetricgroupNoGroup": "TopdownL1;Default", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of time that retirement is stalled due to an L1 miss", + "MetricExpr": "100 * MEM_BOUND_STALLS.LOAD / CPU_CLK_UNHALTED.CORE", + "MetricGroup": "Load_Store_Miss", + "MetricName": "tma_info_bottleneck_%_load_miss_bound_cycles", + "PublicDescription": "Percentage of time that retirement is stalled due to an L1 miss. See Info.Load_Miss_Bound" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to instruction cache misses.", - "MetricExpr": "TOPDOWN_FE_BOUND.ICACHE / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", - "MetricName": "tma_icache_misses", - "MetricThreshold": "tma_icache_misses > 0.05", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of time that retirement is stalled by the Memory Cluster due to a pipeline stall", + "MetricExpr": "100 * LD_HEAD.ANY_AT_RET / CPU_CLK_UNHALTED.CORE", + "MetricGroup": "Mem_Exec", + "MetricName": "tma_info_bottleneck_%_mem_exec_bound_cycles", + "PublicDescription": "Percentage of time that retirement is stalled by the Memory Cluster due to a pipeline stall. See Info.Mem_Exec_Bound" + }, + { + "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)", + "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES", + "MetricName": "tma_info_br_inst_mix_ipbranch" + }, + { + "BriefDescription": "Instruction per (near) call (lower number means higher occurrence rate)", + "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.CALL", + "MetricName": "tma_info_br_inst_mix_ipcall" + }, + { + "BriefDescription": "Instructions per Far Branch ( Far Branches apply upon transition from application to operating system, handling interrupts, exceptions) [lower number means higher occurrence rate]", + "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.FAR_BRANCH:u", + "MetricName": "tma_info_br_inst_mix_ipfarbranch" + }, + { + "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was not taken", + "MetricExpr": "INST_RETIRED.ANY / (BR_MISP_RETIRED.COND - BR_MISP_RETIRED.COND_TAKEN)", + "MetricName": "tma_info_br_inst_mix_ipmisp_cond_ntaken" + }, + { + "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was taken", + "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.COND_TAKEN", + "MetricName": "tma_info_br_inst_mix_ipmisp_cond_taken" + }, + { + "BriefDescription": "Instructions per retired indirect call or jump Branch Misprediction", + "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.INDIRECT", + "MetricName": "tma_info_br_inst_mix_ipmisp_indirect" + }, + { + "BriefDescription": "Instructions per retired return Branch Misprediction", + "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.RETURN", + "MetricName": "tma_info_br_inst_mix_ipmisp_ret" + }, + { + "BriefDescription": "Instructions per retired Branch Misprediction", + "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES", + "MetricName": "tma_info_br_inst_mix_ipmispredict" + }, + { + "BriefDescription": "Ratio of all branches which mispredict", + "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.ALL_BRANCHES", + "MetricName": "tma_info_br_mispredict_bound_branch_mispredict_ratio" + }, + { + "BriefDescription": "Ratio between Mispredicted branches and unknown branches", + "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / BACLEARS.ANY", + "MetricName": "tma_info_br_mispredict_bound_branch_mispredict_to_unknown_branch_ratio" }, { - "BriefDescription": "", - "MetricExpr": "CPU_CLK_UNHALTED.CORE", - "MetricName": "tma_info_core_clks" + "BriefDescription": "Percentage of time that allocation is stalled due to load buffer full", + "MetricExpr": "100 * MEM_SCHEDULER_BLOCK.LD_BUF / CPU_CLK_UNHALTED.CORE", + "MetricName": "tma_info_buffer_stalls_%_load_buffer_stall_cycles" }, { - "BriefDescription": "", - "MetricExpr": "CPU_CLK_UNHALTED.CORE_P", - "MetricName": "tma_info_core_clks_p" + "BriefDescription": "Percentage of time that allocation is stalled due to memory reservation stations full", + "MetricExpr": "100 * MEM_SCHEDULER_BLOCK.RSV / CPU_CLK_UNHALTED.CORE", + "MetricName": "tma_info_buffer_stalls_%_mem_rsv_stall_cycles" + }, + { + "BriefDescription": "Percentage of time that allocation is stalled due to store buffer full", + "MetricExpr": "100 * MEM_SCHEDULER_BLOCK.ST_BUF / CPU_CLK_UNHALTED.CORE", + "MetricName": "tma_info_buffer_stalls_%_store_buffer_stall_cycles" }, { "BriefDescription": "Cycles Per Instruction", - "MetricExpr": "tma_info_core_clks / INST_RETIRED.ANY", + "MetricExpr": "CPU_CLK_UNHALTED.CORE / INST_RETIRED.ANY", "MetricName": "tma_info_core_cpi" }, { "BriefDescription": "Instructions Per Cycle", - "MetricExpr": "INST_RETIRED.ANY / tma_info_core_clks", + "MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.CORE", "MetricName": "tma_info_core_ipc" }, - { - "BriefDescription": "", - "MetricExpr": "5 * tma_info_core_clks", - "MetricName": "tma_info_core_slots" - }, { "BriefDescription": "Uops Per Instruction", "MetricExpr": "UOPS_RETIRED.ALL / INST_RETIRED.ANY", "MetricName": "tma_info_core_upi" }, { - "BriefDescription": "Percent of instruction miss cost that hit in DRAM", + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss hits in the L2", + "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH_L2_HIT / MEM_BOUND_STALLS.IFETCH", + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l2hit" + }, + { + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss hits in the L3", + "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH_LLC_HIT / MEM_BOUND_STALLS.IFETCH", + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l3hit" + }, + { + "BriefDescription": "Percentage of ifetch miss bound stalls, where the ifetch miss subsequently misses in the L3", "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH_DRAM_HIT / MEM_BOUND_STALLS.IFETCH", - "MetricName": "tma_info_frontend_inst_miss_cost_dramhit_percent" + "MetricName": "tma_info_ifetch_miss_bound_%_ifetchmissbound_with_l3miss" }, { - "BriefDescription": "Percent of instruction miss cost that hit in the L2", - "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH_L2_HIT / MEM_BOUND_STALLS.IFETCH", - "MetricName": "tma_info_frontend_inst_miss_cost_l2hit_percent" + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that hit the L2", + "MetricExpr": "100 * MEM_BOUND_STALLS.LOAD_L2_HIT / MEM_BOUND_STALLS.LOAD", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l2hit" }, { - "BriefDescription": "Percent of instruction miss cost that hit in the L3", - "MetricExpr": "100 * MEM_BOUND_STALLS.IFETCH_LLC_HIT / MEM_BOUND_STALLS.IFETCH", - "MetricName": "tma_info_frontend_inst_miss_cost_l3hit_percent" + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that hit the L3", + "MetricExpr": "100 * MEM_BOUND_STALLS.LOAD_LLC_HIT / MEM_BOUND_STALLS.LOAD", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l3hit" }, { - "BriefDescription": "Ratio of all branches which mispredict", - "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_branch_mispredict_ratio" + "BriefDescription": "Percentage of memory bound stalls where retirement is stalled due to an L1 miss that subsequently misses the L3", + "MetricExpr": "100 * MEM_BOUND_STALLS.LOAD_DRAM_HIT / MEM_BOUND_STALLS.LOAD", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_miss_bound_%_loadmissbound_with_l3miss" }, { - "BriefDescription": "Ratio between Mispredicted branches and unknown branches", - "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / BACLEARS.ANY", - "MetricName": "tma_info_inst_mix_branch_mispredict_to_unknown_branch_ratio" + "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a pipeline block", + "MetricExpr": "100 * LD_HEAD.L1_BOUND_AT_RET / CPU_CLK_UNHALTED.CORE", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_l1_bound" }, { - "BriefDescription": "Percentage of all uops which are FPDiv uops", - "MetricExpr": "100 * UOPS_RETIRED.FPDIV / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_fpdiv_uop_ratio" + "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement", + "MetricExpr": "100 * (LD_HEAD.L1_BOUND_AT_RET + MEM_BOUND_STALLS.LOAD) / CPU_CLK_UNHALTED.CORE", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_load_bound" }, { - "BriefDescription": "Percentage of all uops which are IDiv uops", - "MetricExpr": "100 * UOPS_RETIRED.IDIV / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_idiv_uop_ratio" + "BriefDescription": "Counts the number of cycles the core is stalled due to store buffer full", + "MetricExpr": "100 * (MEM_SCHEDULER_BLOCK.ST_BUF / MEM_SCHEDULER_BLOCK.ALL) * tma_mem_scheduler", + "MetricGroup": "load_store_bound", + "MetricName": "tma_info_load_store_bound_store_bound" }, { - "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)", - "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_ipbranch" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory disambiguation", + "MetricExpr": "1e3 * MACHINE_CLEARS.DISAMBIGUATION / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_disamb_pki" }, { - "BriefDescription": "Instruction per (near) call (lower number means higher occurrence rate)", - "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.CALL", - "MetricName": "tma_info_inst_mix_ipcall" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to floating point assists", + "MetricExpr": "1e3 * MACHINE_CLEARS.FP_ASSIST / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_fp_assist_pki" }, { - "BriefDescription": "Instructions per Far Branch", - "MetricExpr": "INST_RETIRED.ANY / (BR_INST_RETIRED.FAR_BRANCH / 2)", - "MetricName": "tma_info_inst_mix_ipfarbranch" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory ordering", + "MetricExpr": "1e3 * MACHINE_CLEARS.MEMORY_ORDERING / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_monuke_pki" }, { - "BriefDescription": "Instructions per Load", - "MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_inst_mix_ipload" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to memory renaming", + "MetricExpr": "1e3 * MACHINE_CLEARS.MRN_NUKE / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_mrn_pki" }, { - "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was not taken", - "MetricExpr": "INST_RETIRED.ANY / (BR_MISP_RETIRED.COND - BR_MISP_RETIRED.COND_TAKEN)", - "MetricName": "tma_info_inst_mix_ipmisp_cond_ntaken" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to page faults", + "MetricExpr": "1e3 * MACHINE_CLEARS.PAGE_FAULT / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_page_fault_pki" }, { - "BriefDescription": "Instructions per retired conditional Branch Misprediction where the branch was taken", - "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.COND_TAKEN", - "MetricName": "tma_info_inst_mix_ipmisp_cond_taken" + "BriefDescription": "Counts the number of machine clears relative to thousands of instructions retired, due to self-modifying code", + "MetricExpr": "1e3 * MACHINE_CLEARS.SMC / INST_RETIRED.ANY", + "MetricName": "tma_info_machine_clear_bound_machine_clears_smc_pki" }, { - "BriefDescription": "Instructions per retired indirect call or jump Branch Misprediction", - "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.INDIRECT", - "MetricName": "tma_info_inst_mix_ipmisp_indirect" + "BriefDescription": "Percentage of total non-speculative loads with an address aliasing block", + "MetricExpr": "100 * LD_BLOCKS.4K_ALIAS / MEM_UOPS_RETIRED.ALL_LOADS", + "MetricName": "tma_info_mem_exec_blocks_%_loads_with_adressaliasing" }, { - "BriefDescription": "Instructions per retired return Branch Misprediction", - "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.RETURN", - "MetricName": "tma_info_inst_mix_ipmisp_ret" + "BriefDescription": "Percentage of total non-speculative loads with a store forward or unknown store address block", + "MetricExpr": "100 * LD_BLOCKS.DATA_UNKNOWN / MEM_UOPS_RETIRED.ALL_LOADS", + "MetricName": "tma_info_mem_exec_blocks_%_loads_with_storefwdblk" }, { - "BriefDescription": "Instructions per retired Branch Misprediction", - "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES", - "MetricName": "tma_info_inst_mix_ipmispredict" + "BriefDescription": "Percentage of Memory Execution Bound due to a first level data cache miss", + "MetricExpr": "100 * LD_HEAD.L1_MISS_AT_RET / LD_HEAD.ANY_AT_RET", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_l1miss" }, { - "BriefDescription": "Instructions per Store", - "MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES", - "MetricName": "tma_info_inst_mix_ipstore" + "BriefDescription": "Percentage of Memory Execution Bound due to other block cases, such as pipeline conflicts, fences, etc", + "MetricExpr": "100 * LD_HEAD.OTHER_AT_RET / LD_HEAD.ANY_AT_RET", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_otherpipelineblks" }, { - "BriefDescription": "Percentage of all uops which are ucode ops", - "MetricExpr": "100 * UOPS_RETIRED.MS / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_microcode_uop_ratio" + "BriefDescription": "Percentage of Memory Execution Bound due to a pagewalk", + "MetricExpr": "100 * LD_HEAD.PGWALK_AT_RET / LD_HEAD.ANY_AT_RET", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_pagewalk" }, { - "BriefDescription": "Percentage of all uops which are x87 uops", - "MetricExpr": "100 * UOPS_RETIRED.X87 / UOPS_RETIRED.ALL", - "MetricName": "tma_info_inst_mix_x87_uop_ratio" + "BriefDescription": "Percentage of Memory Execution Bound due to a second level TLB miss", + "MetricExpr": "100 * LD_HEAD.DTLB_MISS_AT_RET / LD_HEAD.ANY_AT_RET", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_stlbhit" }, { - "BriefDescription": "Percentage of total non-speculative loads with a address aliasing block", - "MetricExpr": "100 * LD_BLOCKS.4K_ALIAS / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_address_alias_blocks" + "BriefDescription": "Percentage of Memory Execution Bound due to a store forward address match", + "MetricExpr": "100 * LD_HEAD.ST_ADDR_AT_RET / LD_HEAD.ANY_AT_RET", + "MetricName": "tma_info_mem_exec_bound_%_loadhead_with_storefwding" }, { - "BriefDescription": "Percentage of total non-speculative loads that are splits", - "MetricExpr": "100 * MEM_UOPS_RETIRED.SPLIT_LOADS / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_load_splits" + "BriefDescription": "Instructions per Load", + "MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS", + "MetricName": "tma_info_mem_mix_ipload" }, { - "BriefDescription": "Percentage of total non-speculative loads with a store forward or unknown store address block", - "MetricExpr": "100 * LD_BLOCKS.DATA_UNKNOWN / MEM_UOPS_RETIRED.ALL_LOADS", - "MetricName": "tma_info_l1_bound_store_fwd_blocks" + "BriefDescription": "Instructions per Store", + "MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES", + "MetricName": "tma_info_mem_mix_ipstore" }, { - "BriefDescription": "Cycle cost per DRAM hit", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_DRAM_HIT / MEM_LOAD_UOPS_RETIRED.DRAM_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_dram_hit" + "BriefDescription": "Percentage of total non-speculative loads that perform one or more locks", + "MetricExpr": "100 * MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_LOADS", + "MetricName": "tma_info_mem_mix_load_locks_ratio" }, { - "BriefDescription": "Cycle cost per L2 hit", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_L2_HIT / MEM_LOAD_UOPS_RETIRED.L2_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_l2_hit" + "BriefDescription": "Percentage of total non-speculative loads that are splits", + "MetricExpr": "100 * MEM_UOPS_RETIRED.SPLIT_LOADS / MEM_UOPS_RETIRED.ALL_LOADS", + "MetricName": "tma_info_mem_mix_load_splits_ratio" }, { - "BriefDescription": "Cycle cost per LLC hit", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_LLC_HIT / MEM_LOAD_UOPS_RETIRED.L3_HIT", - "MetricName": "tma_info_memory_cycles_per_demand_load_l3_hit" + "BriefDescription": "Ratio of mem load uops to all uops", + "MetricExpr": "1e3 * MEM_UOPS_RETIRED.ALL_LOADS / UOPS_RETIRED.ALL", + "MetricName": "tma_info_mem_mix_memload_ratio" }, { - "BriefDescription": "load ops retired per 1000 instruction", - "MetricExpr": "1e3 * MEM_UOPS_RETIRED.ALL_LOADS / INST_RETIRED.ANY", - "MetricName": "tma_info_memory_memloadpki" + "BriefDescription": "Percentage of time that the core is stalled due to a TPAUSE or UMWAIT instruction", + "MetricExpr": "100 * SERIALIZATION.C01_MS_SCB / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricName": "tma_info_serialization _%_tpause_cycles" }, { "BriefDescription": "Average CPU Utilization",
@@ -428,194 +468,122 @@ }, { "BriefDescription": "Fraction of cycles spent in Kernel mode", - "MetricExpr": "cpu@CPU_CLK_UNHALTED.CORE@k / CPU_CLK_UNHALTED.CORE", + "MetricExpr": "cpu@CPU_CLK_UNHALTED.CORE_P@k / CPU_CLK_UNHALTED.CORE", "MetricGroup": "Summary", "MetricName": "tma_info_system_kernel_utilization" }, { "BriefDescription": "Average Frequency Utilization relative nominal frequency", - "MetricExpr": "tma_info_core_clks / CPU_CLK_UNHALTED.REF_TSC", + "MetricExpr": "CPU_CLK_UNHALTED.CORE / CPU_CLK_UNHALTED.REF_TSC", "MetricGroup": "Power", "MetricName": "tma_info_system_turbo_utilization" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to Instruction Table Lookaside Buffer (ITLB) misses.", - "MetricExpr": "TOPDOWN_FE_BOUND.ITLB / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_latency_group", - "MetricName": "tma_itlb_misses", - "MetricThreshold": "tma_itlb_misses > 0.05", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of all uops which are FPDiv uops", + "MetricExpr": "100 * UOPS_RETIRED.FPDIV / UOPS_RETIRED.ALL", + "MetricName": "tma_info_uop_mix_fpdiv_uop_ratio" }, { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a load block.", - "MetricExpr": "LD_HEAD.L1_BOUND_AT_RET / tma_info_core_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l1_bound", - "MetricThreshold": "tma_l1_bound > 0.1", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of all uops which are IDiv uops", + "MetricExpr": "100 * UOPS_RETIRED.IDIV / UOPS_RETIRED.ALL", + "MetricName": "tma_info_uop_mix_idiv_uop_ratio" }, { - "BriefDescription": "Counts the number of cycles a core is stalled due to a demand load which hit in the L2 Cache.", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_L2_HIT / tma_info_core_clks - max((MEM_BOUND_STALLS.LOAD - LD_HEAD.L1_MISS_AT_RET) / tma_info_core_clks, 0) * MEM_BOUND_STALLS.LOAD_L2_HIT / MEM_BOUND_STALLS.LOAD", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l2_bound", - "MetricThreshold": "tma_l2_bound > 0.1", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of all uops which are microcode ops", + "MetricExpr": "100 * UOPS_RETIRED.MS / UOPS_RETIRED.ALL", + "MetricName": "tma_info_uop_mix_microcode_uop_ratio" }, { - "BriefDescription": "Counts the number of cycles a core is stalled due to a demand load which hit in the Last Level Cache (LLC) or other core with HITE/F/M.", - "MetricExpr": "MEM_BOUND_STALLS.LOAD_LLC_HIT / tma_info_core_clks - max((MEM_BOUND_STALLS.LOAD - LD_HEAD.L1_MISS_AT_RET) / tma_info_core_clks, 0) * MEM_BOUND_STALLS.LOAD_LLC_HIT / MEM_BOUND_STALLS.LOAD", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_l3_bound", - "MetricThreshold": "tma_l3_bound > 0.1", - "ScaleUnit": "100%" + "BriefDescription": "Percentage of all uops which are x87 uops", + "MetricExpr": "100 * UOPS_RETIRED.X87 / UOPS_RETIRED.ALL", + "MetricName": "tma_info_uop_mix_x87_uop_ratio" }, { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to load buffer full", - "MetricExpr": "tma_mem_scheduler * MEM_SCHEDULER_BLOCK.LD_BUF / MEM_SCHEDULER_BLOCK.ALL", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_ld_buffer", - "MetricThreshold": "tma_ld_buffer > 0.05", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to Instruction Table Lookaside Buffer (ITLB) misses.", + "MetricExpr": "TOPDOWN_FE_BOUND.ITLB / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_latency_group", + "MetricName": "tma_itlb_misses", + "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_ifetch_latency > 0.15 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a machine clear (nuke) of any kind including memory ordering and memory disambiguation.", - "MetricExpr": "TOPDOWN_BAD_SPECULATION.MACHINE_CLEARS / tma_info_core_slots", + "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend because allocation is stalled due to a machine clear (nuke) of any kind including memory ordering and memory disambiguation", + "MetricExpr": "TOPDOWN_BAD_SPECULATION.MACHINE_CLEARS / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL2;tma_L2_group;tma_bad_speculation_group", "MetricName": "tma_machine_clears", - "MetricThreshold": "tma_machine_clears > 0.05", + "MetricThreshold": "tma_machine_clears > 0.05 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to memory reservation stalls in which a scheduler is not able to accept uops.", - "MetricExpr": "TOPDOWN_BE_BOUND.MEM_SCHEDULER / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to memory reservation stalls in which a scheduler is not able to accept uops", + "MetricExpr": "TOPDOWN_BE_BOUND.MEM_SCHEDULER / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", "MetricName": "tma_mem_scheduler", - "MetricThreshold": "tma_mem_scheduler > 0.1", + "MetricThreshold": "tma_mem_scheduler > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of cycles the core is stalled due to stores or loads.", - "MetricExpr": "min(tma_backend_bound, LD_HEAD.ANY_AT_RET / tma_info_core_clks + tma_store_bound)", - "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", - "MetricName": "tma_memory_bound", - "MetricThreshold": "tma_memory_bound > 0.2", - "MetricgroupNoGroup": "TopdownL2", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to memory ordering.", - "MetricExpr": "tma_nuke * (MACHINE_CLEARS.MEMORY_ORDERING / MACHINE_CLEARS.SLOW)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_memory_ordering", - "MetricThreshold": "tma_memory_ordering > 0.02", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of uops that are from the complex flows issued by the micro-sequencer (MS)", - "MetricExpr": "UOPS_RETIRED.MS / tma_info_core_slots", - "MetricGroup": "TopdownL2;tma_L2_group;tma_retiring_group", - "MetricName": "tma_ms_uops", - "MetricThreshold": "tma_ms_uops > 0.05", - "MetricgroupNoGroup": "TopdownL2", - "PublicDescription": "Counts the number of uops that are from the complex flows issued by the micro-sequencer (MS). This includes uops from flows due to complex instructions, faults, assists, and inserted flows.", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to IEC or FPC RAT stalls, which can be due to FIQ or IEC reservation stalls in which the integer, floating point or SIMD scheduler is not able to accept uops.", - "MetricExpr": "TOPDOWN_BE_BOUND.NON_MEM_SCHEDULER / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to IEC or FPC RAT stalls, which can be due to FIQ or IEC reservation stalls in which the integer, floating point or SIMD scheduler is not able to accept uops", + "MetricExpr": "TOPDOWN_BE_BOUND.NON_MEM_SCHEDULER / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", "MetricName": "tma_non_mem_scheduler", - "MetricThreshold": "tma_non_mem_scheduler > 0.1", + "MetricThreshold": "tma_non_mem_scheduler > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear (slow nuke).", - "MetricExpr": "TOPDOWN_BAD_SPECULATION.NUKE / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to a machine clear that requires the use of microcode (slow nuke)", + "MetricExpr": "TOPDOWN_BAD_SPECULATION.NUKE / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_machine_clears_group", "MetricName": "tma_nuke", - "MetricThreshold": "tma_nuke > 0.05", + "MetricThreshold": "tma_nuke > 0.05 & (tma_machine_clears > 0.05 & tma_bad_speculation > 0.15)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to other common frontend stalls not categorized.", - "MetricExpr": "TOPDOWN_FE_BOUND.OTHER / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to other common frontend stalls not categorized.", + "MetricExpr": "TOPDOWN_FE_BOUND.OTHER / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_other_fb", - "MetricThreshold": "tma_other_fb > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a number of other load blocks.", - "MetricExpr": "LD_HEAD.OTHER_AT_RET / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_other_l1", - "MetricThreshold": "tma_other_l1 > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load miss which hits in the L2, LLC, DRAM or MMIO (Non-DRAM) but could not be correctly attributed or cycles in which the load miss is waiting on a request buffer.", - "MetricExpr": "max(0, tma_memory_bound - (tma_store_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_dram_bound))", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_other_load_store", - "MetricThreshold": "tma_other_load_store > 0.1", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of uops retired excluding ms and fp div uops.", - "MetricExpr": "(TOPDOWN_RETIRING.ALL - UOPS_RETIRED.MS - UOPS_RETIRED.FPDIV) / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_base_group", - "MetricName": "tma_other_ret", - "MetricThreshold": "tma_other_ret > 0.3", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to page faults.", - "MetricExpr": "tma_nuke * (MACHINE_CLEARS.PAGE_FAULT / MACHINE_CLEARS.SLOW)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_page_fault", - "MetricThreshold": "tma_page_fault > 0.02", + "MetricThreshold": "tma_other_fb > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to wrong predecodes.", - "MetricExpr": "TOPDOWN_FE_BOUND.PREDECODE / tma_info_core_slots", - "MetricGroup": "TopdownL3;tma_L3_group;tma_fetch_bandwidth_group", + "BriefDescription": "Counts the number of issue slots that were not delivered by the frontend due to wrong predecodes.", + "MetricExpr": "TOPDOWN_FE_BOUND.PREDECODE / (5 * CPU_CLK_UNHALTED.CORE)", + "MetricGroup": "TopdownL3;tma_L3_group;tma_ifetch_bandwidth_group", "MetricName": "tma_predecode", - "MetricThreshold": "tma_predecode > 0.05", + "MetricThreshold": "tma_predecode > 0.05 & (tma_ifetch_bandwidth > 0.1 & tma_frontend_bound > 0.2)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the physical register file unable to accept an entry (marble stalls).", - "MetricExpr": "TOPDOWN_BE_BOUND.REGISTER / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the physical register file unable to accept an entry (marble stalls)", + "MetricExpr": "TOPDOWN_BE_BOUND.REGISTER / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", "MetricName": "tma_register", - "MetricThreshold": "tma_register > 0.1", + "MetricThreshold": "tma_register > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the reorder buffer being full (ROB stalls).", - "MetricExpr": "TOPDOWN_BE_BOUND.REORDER_BUFFER / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to the reorder buffer being full (ROB stalls)", + "MetricExpr": "TOPDOWN_BE_BOUND.REORDER_BUFFER / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", "MetricName": "tma_reorder_buffer", - "MetricThreshold": "tma_reorder_buffer > 0.1", + "MetricThreshold": "tma_reorder_buffer > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls", - "MetricExpr": "tma_backend_bound", - "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_aux_group", + "BriefDescription": "Counts the number of cycles the core is stalled due to a resource limitation", + "MetricExpr": "tma_backend_bound - tma_core_bound", + "MetricGroup": "TopdownL2;tma_L2_group;tma_backend_bound_group", "MetricName": "tma_resource_bound", - "MetricThreshold": "tma_resource_bound > 0.2", + "MetricThreshold": "tma_resource_bound > 0.2 & tma_backend_bound > 0.1", "MetricgroupNoGroup": "TopdownL2", - "PublicDescription": "Counts the total number of issue slots that were not consumed by the backend due to backend stalls. Note that uops must be available for consumption in order for this event to count. If a uop is not available (IQ is empty), this event will not count.", "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of issue slots that result in retirement slots.", + "BriefDescription": "Counts the number of issue slots that result in retirement slots", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "TOPDOWN_RETIRING.ALL / tma_info_core_slots", + "MetricExpr": "TOPDOWN_RETIRING.ALL / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "Default;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.75",
@@ -623,67 +591,11 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to RSV full relative", - "MetricExpr": "tma_mem_scheduler * MEM_SCHEDULER_BLOCK.RSV / MEM_SCHEDULER_BLOCK.ALL", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_rsv", - "MetricThreshold": "tma_rsv > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to scoreboards from the instruction queue (IQ), jump execution unit (JEU), or microcode sequencer (MS).", - "MetricExpr": "TOPDOWN_BE_BOUND.SERIALIZATION / tma_info_core_slots", + "BriefDescription": "Counts the number of issue slots that were not consumed by the backend due to scoreboards from the instruction queue (IQ), jump execution unit (JEU), or microcode sequencer (MS)", + "MetricExpr": "TOPDOWN_BE_BOUND.SERIALIZATION / (5 * CPU_CLK_UNHALTED.CORE)", "MetricGroup": "TopdownL3;tma_L3_group;tma_resource_bound_group", "MetricName": "tma_serialization", - "MetricThreshold": "tma_serialization > 0.1", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of machine clears relative to the number of nuke slots due to SMC.", - "MetricExpr": "tma_nuke * (MACHINE_CLEARS.SMC / MACHINE_CLEARS.SLOW)", - "MetricGroup": "TopdownL4;tma_L4_group;tma_nuke_group", - "MetricName": "tma_smc", - "MetricThreshold": "tma_smc > 0.02", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles, relative to the number of mem_scheduler slots, in which uops are blocked due to store buffer full", - "MetricExpr": "tma_store_bound", - "MetricGroup": "TopdownL4;tma_L4_group;tma_mem_scheduler_group", - "MetricName": "tma_st_buffer", - "MetricThreshold": "tma_st_buffer > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a first level TLB miss.", - "MetricExpr": "LD_HEAD.DTLB_MISS_AT_RET / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_stlb_hit", - "MetricThreshold": "tma_stlb_hit > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a second level TLB miss requiring a page walk.", - "MetricExpr": "LD_HEAD.PGWALK_AT_RET / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_stlb_miss", - "MetricThreshold": "tma_stlb_miss > 0.05", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles the core is stalled due to store buffer full.", - "MetricExpr": "tma_mem_scheduler * (MEM_SCHEDULER_BLOCK.ST_BUF / MEM_SCHEDULER_BLOCK.ALL)", - "MetricGroup": "TopdownL3;tma_L3_group;tma_memory_bound_group", - "MetricName": "tma_store_bound", - "MetricThreshold": "tma_store_bound > 0.1", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "Counts the number of cycles that the oldest load of the load buffer is stalled at retirement due to a store forward block.", - "MetricExpr": "LD_HEAD.ST_ADDR_AT_RET / tma_info_core_clks", - "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", - "MetricName": "tma_store_fwd_blk", - "MetricThreshold": "tma_store_fwd_blk > 0.05", + "MetricThreshold": "tma_serialization > 0.1 & (tma_resource_bound > 0.2 & tma_backend_bound > 0.1)", "ScaleUnit": "100%" } ]
tools/perf/pmu-events/arch/x86/alderlaken/cache.json +48 -2
270 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlaken/floating-point.json +2 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the number of floating point operations retired that required microcode assist.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.FP_ASSIST", "PublicDescription": "Counts the number of floating point operations retired that required microcode assist, which is not a reflection of the number of FP operations, instructions or uops.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Counts the number of floating point divide uops retired (x87 and SSE, including x87 sqrt).", + "Counter": "0,1,2,3,4,5", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.FPDIV", "PEBS": "1",
tools/perf/pmu-events/arch/x86/alderlaken/frontend.json +3 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the total number of BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xe6", "EventName": "BACLEARS.ANY", "PublicDescription": "Counts the total number of BACLEARS, which occur when the Branch Target Buffer (BTB) prediction or lack thereof, was corrected by a later branch predictor in the frontend. Includes BACLEARS due to all branch types including conditional and unconditional jumps, returns, and indirect branches.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Counts the number of requests to the instruction cache for one or more bytes of a cache line.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x80", "EventName": "ICACHE.ACCESSES", "PublicDescription": "Counts the total number of requests to the instruction cache. The event only counts new cache line accesses, so that multiple back to back fetches to the exact same cache line or byte chunk count as one. Specifically, the event counts when accesses from sequential code crosses the cache line boundary, or when a branch target is moved to a new line or to a non-sequential byte chunk of the same line.",
@@ -17,6 +19,7 @@ }, { "BriefDescription": "Counts the number of instruction cache misses.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x80", "EventName": "ICACHE.MISSES", "PublicDescription": "Counts the number of missed requests to the instruction cache. The event only counts new cache line accesses, so that multiple back to back fetches to the exact same cache line and byte chunk count as one. Specifically, the event counts when accesses from sequential code crosses the cache line boundary, or when a branch target is moved to a new line or to a non-sequential byte chunk of the same line.",
tools/perf/pmu-events/arch/x86/alderlaken/memory.json +11 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer is stalled due to any number of reasons, including an L1 miss, WCB full, pagewalk, store address block or store data block, on a load that retires.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.ANY_AT_RET", "SampleAfterValue": "1000003",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer is stalled due to a core bound stall including a store address match, a DTLB miss or a page walk that detains the load from retiring.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.L1_BOUND_AT_RET", "SampleAfterValue": "1000003",
@@ -15,6 +17,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to a DL1 miss.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.L1_MISS_AT_RET", "SampleAfterValue": "1000003",
@@ -22,6 +25,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to other block cases.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.OTHER_AT_RET", "PublicDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to other block cases such as pipeline conflicts, fences, etc.",
@@ -30,6 +34,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to a pagewalk.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.PGWALK_AT_RET", "SampleAfterValue": "1000003",
@@ -37,6 +42,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to a store address match.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.ST_ADDR_AT_RET", "SampleAfterValue": "1000003",
@@ -44,6 +50,7 @@ }, { "BriefDescription": "Counts the number of machine clears due to memory ordering caused by a snoop from an external agent. Does not count internally generated machine clears such as those due to memory disambiguation.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", "SampleAfterValue": "20003",
@@ -51,6 +58,7 @@ }, { "BriefDescription": "Counts demand data reads that were not supplied by the L3 cache.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_DATA_RD.L3_MISS", "MSRIndex": "0x1a6,0x1a7",
@@ -60,6 +68,7 @@ }, { "BriefDescription": "Counts demand data reads that were not supplied by the L3 cache. [L3_MISS_LOCAL is alias to L3_MISS]", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_DATA_RD.L3_MISS_LOCAL", "MSRIndex": "0x1a6,0x1a7",
@@ -69,6 +78,7 @@ }, { "BriefDescription": "Counts demand reads for ownership (RFO) and software prefetches for exclusive ownership (PREFETCHW) that were not supplied by the L3 cache.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_RFO.L3_MISS", "MSRIndex": "0x1a6,0x1a7",
@@ -78,6 +88,7 @@ }, { "BriefDescription": "Counts demand reads for ownership (RFO) and software prefetches for exclusive ownership (PREFETCHW) that were not supplied by the L3 cache. [L3_MISS_LOCAL is alias to L3_MISS]", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_RFO.L3_MISS_LOCAL", "MSRIndex": "0x1a6,0x1a7",
tools/perf/pmu-events/arch/x86/alderlaken/metricgroups.json +9 -12
@@ -1,26 +1,23 @@ { + "Flops": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "Ifetch": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "Load_Store_Miss": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "Mem_Exec": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Power": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Summary": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "TopdownL1": "Metrics for top-down breakdown at level 1", "TopdownL2": "Metrics for top-down breakdown at level 2", "TopdownL3": "Metrics for top-down breakdown at level 3", - "TopdownL4": "Metrics for top-down breakdown at level 4", + "load_store_bound": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "tma_L1_group": "Metrics for top-down breakdown at level 1", "tma_L2_group": "Metrics for top-down breakdown at level 2", "tma_L3_group": "Metrics for top-down breakdown at level 3", - "tma_L4_group": "Metrics for top-down breakdown at level 4", - "tma_backend_bound_aux_group": "Metrics contributing to tma_backend_bound_aux category", "tma_backend_bound_group": "Metrics contributing to tma_backend_bound category", "tma_bad_speculation_group": "Metrics contributing to tma_bad_speculation category", - "tma_base_group": "Metrics contributing to tma_base category", - "tma_fetch_bandwidth_group": "Metrics contributing to tma_fetch_bandwidth category", - "tma_fetch_latency_group": "Metrics contributing to tma_fetch_latency category", + "tma_core_bound_group": "Metrics contributing to tma_core_bound category", "tma_frontend_bound_group": "Metrics contributing to tma_frontend_bound category", - "tma_l1_bound_group": "Metrics contributing to tma_l1_bound category", + "tma_ifetch_bandwidth_group": "Metrics contributing to tma_ifetch_bandwidth category", + "tma_ifetch_latency_group": "Metrics contributing to tma_ifetch_latency category", "tma_machine_clears_group": "Metrics contributing to tma_machine_clears category", - "tma_mem_scheduler_group": "Metrics contributing to tma_mem_scheduler category", - "tma_memory_bound_group": "Metrics contributing to tma_memory_bound category", - "tma_nuke_group": "Metrics contributing to tma_nuke category", - "tma_resource_bound_group": "Metrics contributing to tma_resource_bound category", - "tma_retiring_group": "Metrics contributing to tma_retiring category" + "tma_resource_bound_group": "Metrics contributing to tma_resource_bound category" }
tools/perf/pmu-events/arch/x86/alderlaken/other.json +13 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "This event is deprecated. [This event is alias to MISC_RETIRED.LBR_INSERTS]", + "Counter": "0,1,2,3,4,5", "Deprecated": "1", "EventCode": "0xe4", "EventName": "LBR_INSERTS.ANY",
@@ -10,6 +11,7 @@ }, { "BriefDescription": "Counts modified writebacks from L1 cache and L2 cache that have any type of response.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.COREWB_M.ANY_RESPONSE", "MSRIndex": "0x1a6,0x1a7",
@@ -19,6 +21,7 @@ }, { "BriefDescription": "Counts demand data reads that have any type of response.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_DATA_RD.ANY_RESPONSE", "MSRIndex": "0x1a6,0x1a7",
@@ -28,6 +31,7 @@ }, { "BriefDescription": "Counts demand reads for ownership (RFO) and software prefetches for exclusive ownership (PREFETCHW) that have any type of response.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.DEMAND_RFO.ANY_RESPONSE", "MSRIndex": "0x1a6,0x1a7",
@@ -37,11 +41,20 @@ }, { "BriefDescription": "Counts streaming stores that have any type of response.", + "Counter": "0,1,2,3,4,5", "EventCode": "0xB7", "EventName": "OCR.STREAMING_WR.ANY_RESPONSE", "MSRIndex": "0x1a6,0x1a7", "MSRValue": "0x10800", "SampleAfterValue": "100003", "UMask": "0x1" + }, + { + "BriefDescription": "Counts the number of issue slots in a UMWAIT or TPAUSE instruction where no uop issues due to the instruction putting the CPU into the C0.1 activity state. For Tremont, UMWAIT and TPAUSE will only put the CPU into C0.1 activity state (not C0.2 activity state)", + "Counter": "0,1,2,3,4,5", + "EventCode": "0x75", + "EventName": "SERIALIZATION.C01_MS_SCB", + "SampleAfterValue": "200003", + "UMask": "0x4" } ]
tools/perf/pmu-events/arch/x86/alderlaken/pipeline.json +80 -0
513 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/alderlaken/uncore-interconnect.json +19 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Number of requests allocated in Coherency Tracker.", + "Counter": "0,1", "EventCode": "0x84", "EventName": "UNC_ARB_COH_TRK_REQUESTS.ALL", "PerPkg": "1",
@@ -9,56 +10,69 @@ }, { "BriefDescription": "Each cycle counts number of any coherent request at memory controller that were issued by any core.", + "Counter": "0", "EventCode": "0x85", "EventName": "UNC_ARB_DAT_OCCUPANCY.ALL", + "Experimental": "1", "PerPkg": "1", "UMask": "0x1", "Unit": "ARB" }, { "BriefDescription": "Each cycle counts number of coherent reads pending on data return from memory controller that were issued by any core.", + "Counter": "0", "EventCode": "0x85", "EventName": "UNC_ARB_DAT_OCCUPANCY.RD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" }, { "BriefDescription": "This event is deprecated. Refer to new event UNC_ARB_REQ_TRK_REQUEST.DRD", + "Counter": "0,1", "Deprecated": "1", "EventCode": "0x81", "EventName": "UNC_ARB_DAT_REQUESTS.RD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" }, { "BriefDescription": "This event is deprecated. Refer to new event UNC_ARB_DAT_OCCUPANCY.ALL", + "Counter": "0", "Deprecated": "1", "EventCode": "0x85", "EventName": "UNC_ARB_IFA_OCCUPANCY.ALL", + "Experimental": "1", "PerPkg": "1", "UMask": "0x1", "Unit": "ARB" }, { "BriefDescription": "Each cycle count number of 'valid' coherent Data Read entries . Such entry is defined as valid when it is allocated till deallocation. Doesn't include prefetches [This event is alias to UNC_ARB_TRK_OCCUPANCY.RD]", + "Counter": "0", "EventCode": "0x80", "EventName": "UNC_ARB_REQ_TRK_OCCUPANCY.DRD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" }, { "BriefDescription": "Number of all coherent Data Read entries. Doesn't include prefetches [This event is alias to UNC_ARB_TRK_REQUESTS.RD]", + "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_REQ_TRK_REQUEST.DRD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" }, { "BriefDescription": "Each cycle counts number of all outgoing valid entries in ReqTrk. Such entry is defined as valid from its allocation in ReqTrk till deallocation. Accounts for Coherent and non-coherent traffic.", + "Counter": "0", "EventCode": "0x80", "EventName": "UNC_ARB_TRK_OCCUPANCY.ALL", "PerPkg": "1",
@@ -67,14 +81,17 @@ }, { "BriefDescription": "Each cycle count number of 'valid' coherent Data Read entries . Such entry is defined as valid when it is allocated till deallocation. Doesn't include prefetches [This event is alias to UNC_ARB_REQ_TRK_OCCUPANCY.DRD]", + "Counter": "0", "EventCode": "0x80", "EventName": "UNC_ARB_TRK_OCCUPANCY.RD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" }, { "BriefDescription": "Counts the number of coherent and in-coherent requests initiated by IA cores, processor graphic units, or LLC.", + "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_TRK_REQUESTS.ALL", "PerPkg": "1",
@@ -83,8 +100,10 @@ }, { "BriefDescription": "Number of all coherent Data Read entries. Doesn't include prefetches [This event is alias to UNC_ARB_REQ_TRK_REQUEST.DRD]", + "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_TRK_REQUESTS.RD", + "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB"
tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json +25 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).", + "Counter": "0", "EventCode": "0xff", "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN", "PerPkg": "1",
@@ -10,6 +11,7 @@ }, { "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.", + "Counter": "1", "EventCode": "0xff", "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN", "PerPkg": "1",
@@ -18,6 +20,7 @@ }, { "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).", + "Counter": "3", "EventCode": "0xff", "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN", "PerPkg": "1",
@@ -27,6 +30,7 @@ }, { "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.", + "Counter": "4", "EventCode": "0xff", "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN", "PerPkg": "1",
@@ -35,6 +39,7 @@ }, { "BriefDescription": "ACT command for a read request sent to DRAM", + "Counter": "0,1,2,3,4", "EventCode": "0x24", "EventName": "UNC_M_ACT_COUNT_RD", "PerPkg": "1",
@@ -42,6 +47,7 @@ }, { "BriefDescription": "ACT command sent to DRAM", + "Counter": "0,1,2,3,4", "EventCode": "0x26", "EventName": "UNC_M_ACT_COUNT_TOTAL", "PerPkg": "1",
@@ -49,6 +55,7 @@ }, { "BriefDescription": "ACT command for a write request sent to DRAM", + "Counter": "0,1,2,3,4", "EventCode": "0x25", "EventName": "UNC_M_ACT_COUNT_WR", "PerPkg": "1",
@@ -56,6 +63,7 @@ }, { "BriefDescription": "Read CAS command sent to DRAM", + "Counter": "0,1,2,3,4", "EventCode": "0x22", "EventName": "UNC_M_CAS_COUNT_RD", "PerPkg": "1",
@@ -63,6 +71,7 @@ }, { "BriefDescription": "Write CAS command sent to DRAM", + "Counter": "0,1,2,3,4", "EventCode": "0x23", "EventName": "UNC_M_CAS_COUNT_WR", "PerPkg": "1",
@@ -70,6 +79,7 @@ }, { "BriefDescription": "Number of clocks", + "Counter": "0,1,2,3,4", "EventCode": "0x01", "EventName": "UNC_M_CLOCKTICKS", "PerPkg": "1",
@@ -77,6 +87,7 @@ }, { "BriefDescription": "incoming read request page status is Page Empty", + "Counter": "0,1,2,3,4", "EventCode": "0x1D", "EventName": "UNC_M_DRAM_PAGE_EMPTY_RD", "PerPkg": "1",
@@ -84,6 +95,7 @@ }, { "BriefDescription": "incoming write request page status is Page Empty", + "Counter": "0,1,2,3,4", "EventCode": "0x20", "EventName": "UNC_M_DRAM_PAGE_EMPTY_WR", "PerPkg": "1",
@@ -91,6 +103,7 @@ }, { "BriefDescription": "incoming read request page status is Page Hit", + "Counter": "0,1,2,3,4", "EventCode": "0x1C", "EventName": "UNC_M_DRAM_PAGE_HIT_RD", "PerPkg": "1",
@@ -98,6 +111,7 @@ }, { "BriefDescription": "incoming write request page status is Page Hit", + "Counter": "0,1,2,3,4", "EventCode": "0x1F", "EventName": "UNC_M_DRAM_PAGE_HIT_WR", "PerPkg": "1",
@@ -105,6 +119,7 @@ }, { "BriefDescription": "incoming read request page status is Page Miss", + "Counter": "0,1,2,3,4", "EventCode": "0x1E", "EventName": "UNC_M_DRAM_PAGE_MISS_RD", "PerPkg": "1",
@@ -112,6 +127,7 @@ }, { "BriefDescription": "incoming write request page status is Page Miss", + "Counter": "0,1,2,3,4", "EventCode": "0x21", "EventName": "UNC_M_DRAM_PAGE_MISS_WR", "PerPkg": "1",
@@ -119,6 +135,7 @@ }, { "BriefDescription": "Any Rank at Hot state", + "Counter": "0,1,2,3,4", "EventCode": "0x19", "EventName": "UNC_M_DRAM_THERMAL_HOT", "PerPkg": "1",
@@ -126,6 +143,7 @@ }, { "BriefDescription": "Any Rank at Warm state", + "Counter": "0,1,2,3,4", "EventCode": "0x1A", "EventName": "UNC_M_DRAM_THERMAL_WARM", "PerPkg": "1",
@@ -133,6 +151,7 @@ }, { "BriefDescription": "Incoming read prefetch request from IA.", + "Counter": "0,1,2,3,4", "EventCode": "0x0A", "EventName": "UNC_M_PREFETCH_RD", "PerPkg": "1",
@@ -140,6 +159,7 @@ }, { "BriefDescription": "PRE command sent to DRAM due to page table idle timer expiration", + "Counter": "0,1,2,3,4", "EventCode": "0x28", "EventName": "UNC_M_PRE_COUNT_IDLE", "PerPkg": "1",
@@ -147,6 +167,7 @@ }, { "BriefDescription": "PRE command sent to DRAM for a read/write request", + "Counter": "0,1,2,3,4", "EventCode": "0x27", "EventName": "UNC_M_PRE_COUNT_PAGE_MISS", "PerPkg": "1",
@@ -154,6 +175,7 @@ }, { "BriefDescription": "Incoming VC0 read request", + "Counter": "0,1,2,3,4", "EventCode": "0x02", "EventName": "UNC_M_VC0_REQUESTS_RD", "PerPkg": "1",
@@ -161,6 +183,7 @@ }, { "BriefDescription": "Incoming VC0 write request", + "Counter": "0,1,2,3,4", "EventCode": "0x03", "EventName": "UNC_M_VC0_REQUESTS_WR", "PerPkg": "1",
@@ -168,6 +191,7 @@ }, { "BriefDescription": "Incoming VC1 read request", + "Counter": "0,1,2,3,4", "EventCode": "0x04", "EventName": "UNC_M_VC1_REQUESTS_RD", "PerPkg": "1",
@@ -175,6 +199,7 @@ }, { "BriefDescription": "Incoming VC1 write request", + "Counter": "0,1,2,3,4", "EventCode": "0x05", "EventName": "UNC_M_VC1_REQUESTS_WR", "PerPkg": "1",
tools/perf/pmu-events/arch/x86/alderlaken/uncore-other.json +1 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "This 48-bit fixed counter counts the UCLK cycles.", + "Counter": "FIXED", "EventCode": "0xff", "EventName": "UNC_CLOCK.SOCKET", "PerPkg": "1",
tools/perf/pmu-events/arch/x86/alderlaken/virtual-memory.json +6 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the number of page walks completed due to load DTLB misses to any page size.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x08", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", "PublicDescription": "Counts the number of page walks completed due to loads (including SW prefetches) whose address translations missed in all Translation Lookaside Buffer (TLB) levels and were mapped to any page size. Includes page walks that page fault.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Counts the number of page walks completed due to store DTLB misses to any page size.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x49", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", "PublicDescription": "Counts the number of page walks completed due to stores whose address translations missed in all Translation Lookaside Buffer (TLB) levels and were mapped to any page size. Includes page walks that page fault.",
@@ -17,6 +19,7 @@ }, { "BriefDescription": "Counts the number of page walks initiated by a instruction fetch that missed the first and second level TLBs.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x85", "EventName": "ITLB_MISSES.MISS_CAUSED_WALK", "SampleAfterValue": "1000003",
@@ -24,6 +27,7 @@ }, { "BriefDescription": "Counts the number of page walks due to an instruction fetch that miss the PDE (Page Directory Entry) cache.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x85", "EventName": "ITLB_MISSES.PDE_CACHE_MISS", "SampleAfterValue": "2000003",
@@ -31,6 +35,7 @@ }, { "BriefDescription": "Counts the number of page walks completed due to instruction fetch misses to any page size.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x85", "EventName": "ITLB_MISSES.WALK_COMPLETED", "PublicDescription": "Counts the number of page walks completed due to instruction fetches whose address translations missed in all Translation Lookaside Buffer (TLB) levels and were mapped to any page size. Includes page walks that page fault.",
@@ -39,6 +44,7 @@ }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to a DTLB miss.", + "Counter": "0,1,2,3,4,5", "EventCode": "0x05", "EventName": "LD_HEAD.DTLB_MISS_AT_RET", "SampleAfterValue": "1000003",
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -1,6 +1,6 @@ Family-model,Version,Filename,EventType GenuineIntel-6-(97|9A|B7|BA|BF),v1.27,alderlake,core -GenuineIntel-6-BE,v1.24,alderlaken,core +GenuineIntel-6-BE,v1.27,alderlaken,core GenuineIntel-6-(1C|26|27|35|36),v5,bonnell,core GenuineIntel-6-(3D|47),v29,broadwell,core GenuineIntel-6-56,v11,broadwellde,core
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Ian Rogers
reviewerPATCH 3/37 perf vendor events: Add bonnell counter information
In reply to Ian RogersAdd counter information necessary for optimizing event grouping the perf tool.
The most recent RFC patch set using this information:
https://lore.kernel.org/lkml/[email protected]/
The information was added in:
https://github.com/intel/perfmon/commit/475892a9690cb048949e593fe39cee65cd4765e1
and later patches.
--- .../pmu-events/arch/x86/bonnell/cache.json | 93 +++++++++++++++++++ .../pmu-events/arch/x86/bonnell/counter.json | 7 ++ .../arch/x86/bonnell/floating-point.json | 32 +++++++ .../pmu-events/arch/x86/bonnell/frontend.json | 11 +++ .../pmu-events/arch/x86/bonnell/memory.json | 19 ++++ .../pmu-events/arch/x86/bonnell/other.json | 56 +++++++++++ .../pmu-events/arch/x86/bonnell/pipeline.json | 44 +++++++++ .../arch/x86/bonnell/virtual-memory.json | 15 +++ 8 files changed, 277 insertions(+) create mode 100644 tools/perf/pmu-events/arch/x86/bonnell/counter.json
tools/perf/pmu-events/arch/x86/bonnell/cache.json +93 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "L1 Data Cacheable reads and writes", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.ALL_CACHE_REF", "SampleAfterValue": "2000000",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "L1 Data reads and writes", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.ALL_REF", "SampleAfterValue": "2000000",
@@ -15,6 +17,7 @@ }, { "BriefDescription": "Modified cache lines evicted from the L1 data cache", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.EVICT", "SampleAfterValue": "200000",
@@ -22,6 +25,7 @@ }, { "BriefDescription": "L1 Cacheable Data Reads", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.LD", "SampleAfterValue": "2000000",
@@ -29,6 +33,7 @@ }, { "BriefDescription": "L1 Data line replacements", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.REPL", "SampleAfterValue": "200000",
@@ -36,6 +41,7 @@ }, { "BriefDescription": "Modified cache lines allocated in the L1 data cache", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.REPLM", "SampleAfterValue": "200000",
@@ -43,6 +49,7 @@ }, { "BriefDescription": "L1 Cacheable Data Writes", + "Counter": "0,1", "EventCode": "0x40", "EventName": "L1D_CACHE.ST", "SampleAfterValue": "2000000",
@@ -50,6 +57,7 @@ }, { "BriefDescription": "Cycles L2 address bus is in use.", + "Counter": "0,1", "EventCode": "0x21", "EventName": "L2_ADS.SELF", "SampleAfterValue": "200000",
@@ -57,6 +65,7 @@ }, { "BriefDescription": "All data requests from the L1 data cache", + "Counter": "0,1", "EventCode": "0x2C", "EventName": "L2_DATA_RQSTS.SELF.E_STATE", "SampleAfterValue": "200000",
@@ -64,6 +73,7 @@ }, { "BriefDescription": "All data requests from the L1 data cache", + "Counter": "0,1", "EventCode": "0x2C", "EventName": "L2_DATA_RQSTS.SELF.I_STATE", "SampleAfterValue": "200000",
@@ -71,6 +81,7 @@ }, { "BriefDescription": "All data requests from the L1 data cache", + "Counter": "0,1", "EventCode": "0x2C", "EventName": "L2_DATA_RQSTS.SELF.MESI", "SampleAfterValue": "200000",
@@ -78,6 +89,7 @@ }, { "BriefDescription": "All data requests from the L1 data cache", + "Counter": "0,1", "EventCode": "0x2C", "EventName": "L2_DATA_RQSTS.SELF.M_STATE", "SampleAfterValue": "200000",
@@ -85,6 +97,7 @@ }, { "BriefDescription": "All data requests from the L1 data cache", + "Counter": "0,1", "EventCode": "0x2C", "EventName": "L2_DATA_RQSTS.SELF.S_STATE", "SampleAfterValue": "200000",
@@ -92,6 +105,7 @@ }, { "BriefDescription": "Cycles the L2 cache data bus is busy.", + "Counter": "0,1", "EventCode": "0x22", "EventName": "L2_DBUS_BUSY.SELF", "SampleAfterValue": "200000",
@@ -99,6 +113,7 @@ }, { "BriefDescription": "Cycles the L2 transfers data to the core.", + "Counter": "0,1", "EventCode": "0x23", "EventName": "L2_DBUS_BUSY_RD.SELF", "SampleAfterValue": "200000",
@@ -106,6 +121,7 @@ }, { "BriefDescription": "L2 cacheable instruction fetch requests", + "Counter": "0,1", "EventCode": "0x28", "EventName": "L2_IFETCH.SELF.E_STATE", "SampleAfterValue": "200000",
@@ -113,6 +129,7 @@ }, { "BriefDescription": "L2 cacheable instruction fetch requests", + "Counter": "0,1", "EventCode": "0x28", "EventName": "L2_IFETCH.SELF.I_STATE", "SampleAfterValue": "200000",
@@ -120,6 +137,7 @@ }, { "BriefDescription": "L2 cacheable instruction fetch requests", + "Counter": "0,1", "EventCode": "0x28", "EventName": "L2_IFETCH.SELF.MESI", "SampleAfterValue": "200000",
@@ -127,6 +145,7 @@ }, { "BriefDescription": "L2 cacheable instruction fetch requests", + "Counter": "0,1", "EventCode": "0x28", "EventName": "L2_IFETCH.SELF.M_STATE", "SampleAfterValue": "200000",
@@ -134,6 +153,7 @@ }, { "BriefDescription": "L2 cacheable instruction fetch requests", + "Counter": "0,1", "EventCode": "0x28", "EventName": "L2_IFETCH.SELF.S_STATE", "SampleAfterValue": "200000",
@@ -141,6 +161,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.ANY.E_STATE", "SampleAfterValue": "200000",
@@ -148,6 +169,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.ANY.I_STATE", "SampleAfterValue": "200000",
@@ -155,6 +177,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.ANY.MESI", "SampleAfterValue": "200000",
@@ -162,6 +185,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.ANY.M_STATE", "SampleAfterValue": "200000",
@@ -169,6 +193,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.ANY.S_STATE", "SampleAfterValue": "200000",
@@ -176,6 +201,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.DEMAND.E_STATE", "SampleAfterValue": "200000",
@@ -183,6 +209,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.DEMAND.I_STATE", "SampleAfterValue": "200000",
@@ -190,6 +217,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.DEMAND.MESI", "SampleAfterValue": "200000",
@@ -197,6 +225,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.DEMAND.M_STATE", "SampleAfterValue": "200000",
@@ -204,6 +233,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.DEMAND.S_STATE", "SampleAfterValue": "200000",
@@ -211,6 +241,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.PREFETCH.E_STATE", "SampleAfterValue": "200000",
@@ -218,6 +249,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.PREFETCH.I_STATE", "SampleAfterValue": "200000",
@@ -225,6 +257,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.PREFETCH.MESI", "SampleAfterValue": "200000",
@@ -232,6 +265,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.PREFETCH.M_STATE", "SampleAfterValue": "200000",
@@ -239,6 +273,7 @@ }, { "BriefDescription": "L2 cache reads", + "Counter": "0,1", "EventCode": "0x29", "EventName": "L2_LD.SELF.PREFETCH.S_STATE", "SampleAfterValue": "200000",
@@ -246,6 +281,7 @@ }, { "BriefDescription": "All read requests from L1 instruction and data caches", + "Counter": "0,1", "EventCode": "0x2D", "EventName": "L2_LD_IFETCH.SELF.E_STATE", "SampleAfterValue": "200000",
@@ -253,6 +289,7 @@ }, { "BriefDescription": "All read requests from L1 instruction and data caches", + "Counter": "0,1", "EventCode": "0x2D", "EventName": "L2_LD_IFETCH.SELF.I_STATE", "SampleAfterValue": "200000",
@@ -260,6 +297,7 @@ }, { "BriefDescription": "All read requests from L1 instruction and data caches", + "Counter": "0,1", "EventCode": "0x2D", "EventName": "L2_LD_IFETCH.SELF.MESI", "SampleAfterValue": "200000",
@@ -267,6 +305,7 @@ }, { "BriefDescription": "All read requests from L1 instruction and data caches", + "Counter": "0,1", "EventCode": "0x2D", "EventName": "L2_LD_IFETCH.SELF.M_STATE", "SampleAfterValue": "200000",
@@ -274,6 +313,7 @@ }, { "BriefDescription": "All read requests from L1 instruction and data caches", + "Counter": "0,1", "EventCode": "0x2D", "EventName": "L2_LD_IFETCH.SELF.S_STATE", "SampleAfterValue": "200000",
@@ -281,6 +321,7 @@ }, { "BriefDescription": "L2 cache misses.", + "Counter": "0,1", "EventCode": "0x24", "EventName": "L2_LINES_IN.SELF.ANY", "SampleAfterValue": "200000",
@@ -288,6 +329,7 @@ }, { "BriefDescription": "L2 cache misses.", + "Counter": "0,1", "EventCode": "0x24", "EventName": "L2_LINES_IN.SELF.DEMAND", "SampleAfterValue": "200000",
@@ -295,6 +337,7 @@ }, { "BriefDescription": "L2 cache misses.", + "Counter": "0,1", "EventCode": "0x24", "EventName": "L2_LINES_IN.SELF.PREFETCH", "SampleAfterValue": "200000",
@@ -302,6 +345,7 @@ }, { "BriefDescription": "L2 cache lines evicted.", + "Counter": "0,1", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SELF.ANY", "SampleAfterValue": "200000",
@@ -309,6 +353,7 @@ }, { "BriefDescription": "L2 cache lines evicted.", + "Counter": "0,1", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SELF.DEMAND", "SampleAfterValue": "200000",
@@ -316,6 +361,7 @@ }, { "BriefDescription": "L2 cache lines evicted.", + "Counter": "0,1", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SELF.PREFETCH", "SampleAfterValue": "200000",
@@ -323,6 +369,7 @@ }, { "BriefDescription": "L2 locked accesses", + "Counter": "0,1", "EventCode": "0x2B", "EventName": "L2_LOCK.SELF.E_STATE", "SampleAfterValue": "200000",
@@ -330,6 +377,7 @@ }, { "BriefDescription": "L2 locked accesses", + "Counter": "0,1", "EventCode": "0x2B", "EventName": "L2_LOCK.SELF.I_STATE", "SampleAfterValue": "200000",
@@ -337,6 +385,7 @@ }, { "BriefDescription": "L2 locked accesses", + "Counter": "0,1", "EventCode": "0x2B", "EventName": "L2_LOCK.SELF.MESI", "SampleAfterValue": "200000",
@@ -344,6 +393,7 @@ }, { "BriefDescription": "L2 locked accesses", + "Counter": "0,1", "EventCode": "0x2B", "EventName": "L2_LOCK.SELF.M_STATE", "SampleAfterValue": "200000",
@@ -351,6 +401,7 @@ }, { "BriefDescription": "L2 locked accesses", + "Counter": "0,1", "EventCode": "0x2B", "EventName": "L2_LOCK.SELF.S_STATE", "SampleAfterValue": "200000",
@@ -358,6 +409,7 @@ }, { "BriefDescription": "L2 cache line modifications.", + "Counter": "0,1", "EventCode": "0x25", "EventName": "L2_M_LINES_IN.SELF", "SampleAfterValue": "200000",
@@ -365,6 +417,7 @@ }, { "BriefDescription": "Modified lines evicted from the L2 cache", + "Counter": "0,1", "EventCode": "0x27", "EventName": "L2_M_LINES_OUT.SELF.ANY", "SampleAfterValue": "200000",
@@ -372,6 +425,7 @@ }, { "BriefDescription": "Modified lines evicted from the L2 cache", + "Counter": "0,1", "EventCode": "0x27", "EventName": "L2_M_LINES_OUT.SELF.DEMAND", "SampleAfterValue": "200000",
@@ -379,6 +433,7 @@ }, { "BriefDescription": "Modified lines evicted from the L2 cache", + "Counter": "0,1", "EventCode": "0x27", "EventName": "L2_M_LINES_OUT.SELF.PREFETCH", "SampleAfterValue": "200000",
@@ -386,6 +441,7 @@ }, { "BriefDescription": "Cycles no L2 cache requests are pending", + "Counter": "0,1", "EventCode": "0x32", "EventName": "L2_NO_REQ.SELF", "SampleAfterValue": "200000",
@@ -393,6 +449,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.ANY.E_STATE", "SampleAfterValue": "200000",
@@ -400,6 +457,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.ANY.I_STATE", "SampleAfterValue": "200000",
@@ -407,6 +465,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.ANY.MESI", "SampleAfterValue": "200000",
@@ -414,6 +473,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.ANY.M_STATE", "SampleAfterValue": "200000",
@@ -421,6 +481,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.ANY.S_STATE", "SampleAfterValue": "200000",
@@ -428,6 +489,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.DEMAND.E_STATE", "SampleAfterValue": "200000",
@@ -435,6 +497,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.DEMAND.I_STATE", "SampleAfterValue": "200000",
@@ -442,6 +505,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.DEMAND.MESI", "SampleAfterValue": "200000",
@@ -449,6 +513,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.DEMAND.M_STATE", "SampleAfterValue": "200000",
@@ -456,6 +521,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.DEMAND.S_STATE", "SampleAfterValue": "200000",
@@ -463,6 +529,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.PREFETCH.E_STATE", "SampleAfterValue": "200000",
@@ -470,6 +537,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.PREFETCH.I_STATE", "SampleAfterValue": "200000",
@@ -477,6 +545,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.PREFETCH.MESI", "SampleAfterValue": "200000",
@@ -484,6 +553,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.PREFETCH.M_STATE", "SampleAfterValue": "200000",
@@ -491,6 +561,7 @@ }, { "BriefDescription": "Rejected L2 cache requests", + "Counter": "0,1", "EventCode": "0x30", "EventName": "L2_REJECT_BUSQ.SELF.PREFETCH.S_STATE", "SampleAfterValue": "200000",
@@ -498,6 +569,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.ANY.E_STATE", "SampleAfterValue": "200000",
@@ -505,6 +577,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.ANY.I_STATE", "SampleAfterValue": "200000",
@@ -512,6 +585,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.ANY.MESI", "SampleAfterValue": "200000",
@@ -519,6 +593,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.ANY.M_STATE", "SampleAfterValue": "200000",
@@ -526,6 +601,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.ANY.S_STATE", "SampleAfterValue": "200000",
@@ -533,6 +609,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.DEMAND.E_STATE", "SampleAfterValue": "200000",
@@ -540,6 +617,7 @@ }, { "BriefDescription": "L2 cache demand requests from this core that missed the L2", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.DEMAND.I_STATE", "SampleAfterValue": "200000",
@@ -547,6 +625,7 @@ }, { "BriefDescription": "L2 cache demand requests from this core", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.DEMAND.MESI", "SampleAfterValue": "200000",
@@ -554,6 +633,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.DEMAND.M_STATE", "SampleAfterValue": "200000",
@@ -561,6 +641,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.DEMAND.S_STATE", "SampleAfterValue": "200000",
@@ -568,6 +649,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.PREFETCH.E_STATE", "SampleAfterValue": "200000",
@@ -575,6 +657,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.PREFETCH.I_STATE", "SampleAfterValue": "200000",
@@ -582,6 +665,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.PREFETCH.MESI", "SampleAfterValue": "200000",
@@ -589,6 +673,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.PREFETCH.M_STATE", "SampleAfterValue": "200000",
@@ -596,6 +681,7 @@ }, { "BriefDescription": "L2 cache requests", + "Counter": "0,1", "EventCode": "0x2E", "EventName": "L2_RQSTS.SELF.PREFETCH.S_STATE", "SampleAfterValue": "200000",
@@ -603,6 +689,7 @@ }, { "BriefDescription": "L2 store requests", + "Counter": "0,1", "EventCode": "0x2A", "EventName": "L2_ST.SELF.E_STATE", "SampleAfterValue": "200000",
@@ -610,6 +697,7 @@ }, { "BriefDescription": "L2 store requests", + "Counter": "0,1", "EventCode": "0x2A", "EventName": "L2_ST.SELF.I_STATE", "SampleAfterValue": "200000",
@@ -617,6 +705,7 @@ }, { "BriefDescription": "L2 store requests", + "Counter": "0,1", "EventCode": "0x2A", "EventName": "L2_ST.SELF.MESI", "SampleAfterValue": "200000",
@@ -624,6 +713,7 @@ }, { "BriefDescription": "L2 store requests", + "Counter": "0,1", "EventCode": "0x2A", "EventName": "L2_ST.SELF.M_STATE", "SampleAfterValue": "200000",
@@ -631,6 +721,7 @@ }, { "BriefDescription": "L2 store requests", + "Counter": "0,1", "EventCode": "0x2A", "EventName": "L2_ST.SELF.S_STATE", "SampleAfterValue": "200000",
@@ -638,6 +729,7 @@ }, { "BriefDescription": "Retired loads that hit the L2 cache (precise event).", + "Counter": "0,1", "EventCode": "0xCB", "EventName": "MEM_LOAD_RETIRED.L2_HIT", "SampleAfterValue": "200000",
@@ -645,6 +737,7 @@ }, { "BriefDescription": "Retired loads that miss the L2 cache", + "Counter": "0,1", "EventCode": "0xCB", "EventName": "MEM_LOAD_RETIRED.L2_MISS", "SampleAfterValue": "10000",
tools/perf/pmu-events/arch/x86/bonnell/counter.json new file +7 -0
@@ -0,0 +1,7 @@ +[ + { + "Unit": "core", + "CountersNumFixed": "4", + "CountersNumGeneric": "2" + } +]
tools/perf/pmu-events/arch/x86/bonnell/floating-point.json +32 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Floating point assists for retired operations.", + "Counter": "0,1", "EventCode": "0x11", "EventName": "FP_ASSIST.AR", "SampleAfterValue": "10000",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Floating point assists.", + "Counter": "0,1", "EventCode": "0x11", "EventName": "FP_ASSIST.S", "SampleAfterValue": "10000",
@@ -15,12 +17,14 @@ }, { "BriefDescription": "SIMD assists invoked.", + "Counter": "0,1", "EventCode": "0xCD", "EventName": "SIMD_ASSIST", "SampleAfterValue": "100000" }, { "BriefDescription": "Retired computational Streaming SIMD Extensions (SSE) packed-single instructions.", + "Counter": "0,1", "EventCode": "0xCA", "EventName": "SIMD_COMP_INST_RETIRED.PACKED_SINGLE", "SampleAfterValue": "2000000",
@@ -28,6 +32,7 @@ }, { "BriefDescription": "Retired computational Streaming SIMD Extensions 2 (SSE2) scalar-double instructions.", + "Counter": "0,1", "EventCode": "0xCA", "EventName": "SIMD_COMP_INST_RETIRED.SCALAR_DOUBLE", "SampleAfterValue": "2000000",
@@ -35,6 +40,7 @@ }, { "BriefDescription": "Retired computational Streaming SIMD Extensions (SSE) scalar-single instructions.", + "Counter": "0,1", "EventCode": "0xCA", "EventName": "SIMD_COMP_INST_RETIRED.SCALAR_SINGLE", "SampleAfterValue": "2000000",
@@ -42,12 +48,14 @@ }, { "BriefDescription": "SIMD Instructions retired.", + "Counter": "0,1", "EventCode": "0xCE", "EventName": "SIMD_INSTR_RETIRED", "SampleAfterValue": "2000000" }, { "BriefDescription": "Retired Streaming SIMD Extensions (SSE) packed-single instructions.", + "Counter": "0,1", "EventCode": "0xC7", "EventName": "SIMD_INST_RETIRED.PACKED_SINGLE", "SampleAfterValue": "2000000",
@@ -55,6 +63,7 @@ }, { "BriefDescription": "Retired Streaming SIMD Extensions 2 (SSE2) scalar-double instructions.", + "Counter": "0,1", "EventCode": "0xC7", "EventName": "SIMD_INST_RETIRED.SCALAR_DOUBLE", "SampleAfterValue": "2000000",
@@ -62,6 +71,7 @@ }, { "BriefDescription": "Retired Streaming SIMD Extensions (SSE) scalar-single instructions.", + "Counter": "0,1", "EventCode": "0xC7", "EventName": "SIMD_INST_RETIRED.SCALAR_SINGLE", "SampleAfterValue": "2000000",
@@ -69,6 +79,7 @@ }, { "BriefDescription": "Retired Streaming SIMD Extensions 2 (SSE2) vector instructions.", + "Counter": "0,1", "EventCode": "0xC7", "EventName": "SIMD_INST_RETIRED.VECTOR", "SampleAfterValue": "2000000",
@@ -76,12 +87,14 @@ }, { "BriefDescription": "Saturated arithmetic instructions retired.", + "Counter": "0,1", "EventCode": "0xCF", "EventName": "SIMD_SAT_INSTR_RETIRED", "SampleAfterValue": "2000000" }, { "BriefDescription": "SIMD saturated arithmetic micro-ops retired.", + "Counter": "0,1", "EventCode": "0xB1", "EventName": "SIMD_SAT_UOP_EXEC.AR", "SampleAfterValue": "2000000",
@@ -89,12 +102,14 @@ }, { "BriefDescription": "SIMD saturated arithmetic micro-ops executed.", + "Counter": "0,1", "EventCode": "0xB1", "EventName": "SIMD_SAT_UOP_EXEC.S", "SampleAfterValue": "2000000" }, { "BriefDescription": "SIMD micro-ops retired (excluding stores).", + "Counter": "0,1", "EventCode": "0xB0", "EventName": "SIMD_UOPS_EXEC.AR", "PEBS": "2",
@@ -103,12 +118,14 @@ }, { "BriefDescription": "SIMD micro-ops executed (excluding stores).", + "Counter": "0,1", "EventCode": "0xB0", "EventName": "SIMD_UOPS_EXEC.S", "SampleAfterValue": "2000000" }, { "BriefDescription": "SIMD packed arithmetic micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.ARITHMETIC.AR", "SampleAfterValue": "2000000",
@@ -116,6 +133,7 @@ }, { "BriefDescription": "SIMD packed arithmetic micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.ARITHMETIC.S", "SampleAfterValue": "2000000",
@@ -123,6 +141,7 @@ }, { "BriefDescription": "SIMD packed logical micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.LOGICAL.AR", "SampleAfterValue": "2000000",
@@ -130,6 +149,7 @@ }, { "BriefDescription": "SIMD packed logical micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.LOGICAL.S", "SampleAfterValue": "2000000",
@@ -137,6 +157,7 @@ }, { "BriefDescription": "SIMD packed multiply micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.MUL.AR", "SampleAfterValue": "2000000",
@@ -144,6 +165,7 @@ }, { "BriefDescription": "SIMD packed multiply micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.MUL.S", "SampleAfterValue": "2000000",
@@ -151,6 +173,7 @@ }, { "BriefDescription": "SIMD packed micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.PACK.AR", "SampleAfterValue": "2000000",
@@ -158,6 +181,7 @@ }, { "BriefDescription": "SIMD packed micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.PACK.S", "SampleAfterValue": "2000000",
@@ -165,6 +189,7 @@ }, { "BriefDescription": "SIMD packed shift micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.SHIFT.AR", "SampleAfterValue": "2000000",
@@ -172,6 +197,7 @@ }, { "BriefDescription": "SIMD packed shift micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.SHIFT.S", "SampleAfterValue": "2000000",
@@ -179,6 +205,7 @@ }, { "BriefDescription": "SIMD unpacked micro-ops retired", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.UNPACK.AR", "SampleAfterValue": "2000000",
@@ -186,6 +213,7 @@ }, { "BriefDescription": "SIMD unpacked micro-ops executed", + "Counter": "0,1", "EventCode": "0xB3", "EventName": "SIMD_UOP_TYPE_EXEC.UNPACK.S", "SampleAfterValue": "2000000",
@@ -193,6 +221,7 @@ }, { "BriefDescription": "Floating point computational micro-ops retired.", + "Counter": "0,1", "EventCode": "0x10", "EventName": "X87_COMP_OPS_EXE.ANY.AR", "PEBS": "2",
@@ -201,6 +230,7 @@ }, { "BriefDescription": "Floating point computational micro-ops executed.", + "Counter": "0,1", "EventCode": "0x10", "EventName": "X87_COMP_OPS_EXE.ANY.S", "SampleAfterValue": "2000000",
@@ -208,6 +238,7 @@ }, { "BriefDescription": "FXCH uops retired.", + "Counter": "0,1", "EventCode": "0x10", "EventName": "X87_COMP_OPS_EXE.FXCH.AR", "PEBS": "2",
@@ -216,6 +247,7 @@ }, { "BriefDescription": "FXCH uops executed.", + "Counter": "0,1", "EventCode": "0x10", "EventName": "X87_COMP_OPS_EXE.FXCH.S", "SampleAfterValue": "2000000",
tools/perf/pmu-events/arch/x86/bonnell/frontend.json +11 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "BACLEARS asserted.", + "Counter": "0,1", "EventCode": "0xE6", "EventName": "BACLEARS.ANY", "SampleAfterValue": "2000000",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Cycles during which instruction fetches are stalled.", + "Counter": "0,1", "EventCode": "0x86", "EventName": "CYCLES_ICACHE_MEM_STALLED.ICACHE_MEM_STALLED", "SampleAfterValue": "2000000",
@@ -15,6 +17,7 @@ }, { "BriefDescription": "Decode stall due to IQ full", + "Counter": "0,1", "EventCode": "0x87", "EventName": "DECODE_STALL.IQ_FULL", "SampleAfterValue": "2000000",
@@ -22,6 +25,7 @@ }, { "BriefDescription": "Decode stall due to PFB empty", + "Counter": "0,1", "EventCode": "0x87", "EventName": "DECODE_STALL.PFB_EMPTY", "SampleAfterValue": "2000000",
@@ -29,6 +33,7 @@ }, { "BriefDescription": "Instruction fetches.", + "Counter": "0,1", "EventCode": "0x80", "EventName": "ICACHE.ACCESSES", "SampleAfterValue": "200000",
@@ -36,6 +41,7 @@ }, { "BriefDescription": "Icache hit", + "Counter": "0,1", "EventCode": "0x80", "EventName": "ICACHE.HIT", "SampleAfterValue": "200000",
@@ -43,6 +49,7 @@ }, { "BriefDescription": "Icache miss", + "Counter": "0,1", "EventCode": "0x80", "EventName": "ICACHE.MISSES", "SampleAfterValue": "200000",
@@ -50,6 +57,7 @@ }, { "BriefDescription": "All Instructions decoded", + "Counter": "0,1", "EventCode": "0xAA", "EventName": "MACRO_INSTS.ALL_DECODED", "SampleAfterValue": "2000000",
@@ -57,6 +65,7 @@ }, { "BriefDescription": "CISC macro instructions decoded", + "Counter": "0,1", "EventCode": "0xAA", "EventName": "MACRO_INSTS.CISC_DECODED", "SampleAfterValue": "2000000",
@@ -64,6 +73,7 @@ }, { "BriefDescription": "Non-CISC macro instructions decoded", + "Counter": "0,1", "EventCode": "0xAA", "EventName": "MACRO_INSTS.NON_CISC_DECODED", "SampleAfterValue": "2000000",
@@ -71,6 +81,7 @@ }, { "BriefDescription": "This event counts the cycles where 1 or more uops are issued by the micro-sequencer (MS), including microcode assists and inserted flows, and written to the IQ.", + "Counter": "0,1", "CounterMask": "1", "EventCode": "0xA9", "EventName": "UOPS.MS_CYCLES",
tools/perf/pmu-events/arch/x86/bonnell/memory.json +19 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Nonzero segbase 1 bubble", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.BUBBLE", "SampleAfterValue": "200000",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Nonzero segbase load 1 bubble", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.LD_BUBBLE", "SampleAfterValue": "200000",
@@ -15,6 +17,7 @@ }, { "BriefDescription": "Load splits", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.LD_SPLIT", "SampleAfterValue": "200000",
@@ -22,6 +25,7 @@ }, { "BriefDescription": "Load splits (At Retirement)", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.LD_SPLIT.AR", "SampleAfterValue": "200000",
@@ -29,6 +33,7 @@ }, { "BriefDescription": "Nonzero segbase ld-op-st 1 bubble", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.RMW_BUBBLE", "SampleAfterValue": "200000",
@@ -36,6 +41,7 @@ }, { "BriefDescription": "ld-op-st splits", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.RMW_SPLIT", "SampleAfterValue": "200000",
@@ -43,6 +49,7 @@ }, { "BriefDescription": "Memory references that cross an 8-byte boundary.", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.SPLIT", "SampleAfterValue": "200000",
@@ -50,6 +57,7 @@ }, { "BriefDescription": "Memory references that cross an 8-byte boundary (At Retirement)", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.SPLIT.AR", "SampleAfterValue": "200000",
@@ -57,6 +65,7 @@ }, { "BriefDescription": "Nonzero segbase store 1 bubble", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.ST_BUBBLE", "SampleAfterValue": "200000",
@@ -64,6 +73,7 @@ }, { "BriefDescription": "Store splits", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.ST_SPLIT", "SampleAfterValue": "200000",
@@ -71,6 +81,7 @@ }, { "BriefDescription": "Store splits (Ar Retirement)", + "Counter": "0,1", "EventCode": "0x5", "EventName": "MISALIGN_MEM_REF.ST_SPLIT.AR", "SampleAfterValue": "200000",
@@ -78,6 +89,7 @@ }, { "BriefDescription": "L1 hardware prefetch request", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.HW_PREFETCH", "SampleAfterValue": "2000000",
@@ -85,6 +97,7 @@ }, { "BriefDescription": "Streaming SIMD Extensions (SSE) Prefetch NTA instructions executed", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.PREFETCHNTA", "SampleAfterValue": "200000",
@@ -92,6 +105,7 @@ }, { "BriefDescription": "Streaming SIMD Extensions (SSE) PrefetchT0 instructions executed.", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.PREFETCHT0", "SampleAfterValue": "200000",
@@ -99,6 +113,7 @@ }, { "BriefDescription": "Streaming SIMD Extensions (SSE) PrefetchT1 instructions executed.", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.PREFETCHT1", "SampleAfterValue": "200000",
@@ -106,6 +121,7 @@ }, { "BriefDescription": "Streaming SIMD Extensions (SSE) PrefetchT2 instructions executed.", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.PREFETCHT2", "SampleAfterValue": "200000",
@@ -113,6 +129,7 @@ }, { "BriefDescription": "Any Software prefetch", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.SOFTWARE_PREFETCH", "SampleAfterValue": "200000",
@@ -120,6 +137,7 @@ }, { "BriefDescription": "Any Software prefetch", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.SOFTWARE_PREFETCH.AR", "SampleAfterValue": "200000",
@@ -127,6 +145,7 @@ }, { "BriefDescription": "Streaming SIMD Extensions (SSE) PrefetchT1 and PrefetchT2 instructions executed", + "Counter": "0,1", "EventCode": "0x7", "EventName": "PREFETCH.SW_L2", "SampleAfterValue": "200000",
tools/perf/pmu-events/arch/x86/bonnell/other.json +56 -0
392 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/bonnell/pipeline.json +44 -0
308 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/bonnell/virtual-memory.json +15 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Memory accesses that missed the DTLB.", + "Counter": "0,1", "EventCode": "0x8", "EventName": "DATA_TLB_MISSES.DTLB_MISS", "SampleAfterValue": "200000",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "DTLB misses due to load operations.", + "Counter": "0,1", "EventCode": "0x8", "EventName": "DATA_TLB_MISSES.DTLB_MISS_LD", "SampleAfterValue": "200000",
@@ -15,6 +17,7 @@ }, { "BriefDescription": "DTLB misses due to store operations.", + "Counter": "0,1", "EventCode": "0x8", "EventName": "DATA_TLB_MISSES.DTLB_MISS_ST", "SampleAfterValue": "200000",
@@ -22,6 +25,7 @@ }, { "BriefDescription": "L0 DTLB misses due to load operations.", + "Counter": "0,1", "EventCode": "0x8", "EventName": "DATA_TLB_MISSES.L0_DTLB_MISS_LD", "SampleAfterValue": "200000",
@@ -29,6 +33,7 @@ }, { "BriefDescription": "L0 DTLB misses due to store operations", + "Counter": "0,1", "EventCode": "0x8", "EventName": "DATA_TLB_MISSES.L0_DTLB_MISS_ST", "SampleAfterValue": "200000",
@@ -36,6 +41,7 @@ }, { "BriefDescription": "ITLB flushes.", + "Counter": "0,1", "EventCode": "0x82", "EventName": "ITLB.FLUSH", "SampleAfterValue": "200000",
@@ -43,6 +49,7 @@ }, { "BriefDescription": "ITLB hits.", + "Counter": "0,1", "EventCode": "0x82", "EventName": "ITLB.HIT", "SampleAfterValue": "200000",
@@ -50,6 +57,7 @@ }, { "BriefDescription": "ITLB misses.", + "Counter": "0,1", "EventCode": "0x82", "EventName": "ITLB.MISSES", "PEBS": "2",
@@ -58,6 +66,7 @@ }, { "BriefDescription": "Retired loads that miss the DTLB (precise event).", + "Counter": "0,1", "EventCode": "0xCB", "EventName": "MEM_LOAD_RETIRED.DTLB_MISS", "PEBS": "1",
@@ -66,6 +75,7 @@ }, { "BriefDescription": "Duration of page-walks in core cycles", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.CYCLES", "SampleAfterValue": "2000000",
@@ -73,6 +83,7 @@ }, { "BriefDescription": "Duration of D-side only page walks", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.D_SIDE_CYCLES", "SampleAfterValue": "2000000",
@@ -80,6 +91,7 @@ }, { "BriefDescription": "Number of D-side only page walks", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.D_SIDE_WALKS", "SampleAfterValue": "200000",
@@ -87,6 +99,7 @@ }, { "BriefDescription": "Duration of I-Side page walks", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.I_SIDE_CYCLES", "SampleAfterValue": "2000000",
@@ -94,6 +107,7 @@ }, { "BriefDescription": "Number of I-Side page walks", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.I_SIDE_WALKS", "SampleAfterValue": "200000",
@@ -101,6 +115,7 @@ }, { "BriefDescription": "Number of page-walks executed.", + "Counter": "0,1", "EventCode": "0xC", "EventName": "PAGE_WALKS.WALKS", "SampleAfterValue": "200000",
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Ian Rogers
reviewerPATCH 4/37 perf vendor events: Update broadwell metrics add event counter information
In reply to Ian RogersAdd counter information necessary for optimizing event grouping the perf tool.
The most recent RFC patch set using this information:
https://lore.kernel.org/lkml/[email protected]/
The information was added in:
https://github.com/intel/perfmon/commit/475892a9690cb048949e593fe39cee65cd4765e1
and later patches.
The TMA 4.8 information was updated in:
https://github.com/intel/perfmon/commit/59194d4d90ca50a3fcb2de0d82b9f6fc0c9a5736
--- .../arch/x86/broadwell/bdw-metrics.json | 80 ++--- .../pmu-events/arch/x86/broadwell/cache.json | 275 ++++++++++++++++++ .../arch/x86/broadwell/counter.json | 22 ++ .../arch/x86/broadwell/floating-point.json | 22 ++ .../arch/x86/broadwell/frontend.json | 28 ++ .../pmu-events/arch/x86/broadwell/memory.json | 240 +++++++++++++++ .../arch/x86/broadwell/metricgroups.json | 11 + .../pmu-events/arch/x86/broadwell/other.json | 4 + .../arch/x86/broadwell/pipeline.json | 137 +++++++++ .../arch/x86/broadwell/uncore-cache.json | 24 ++ .../x86/broadwell/uncore-interconnect.json | 7 + .../arch/x86/broadwell/uncore-other.json | 10 - .../arch/x86/broadwell/virtual-memory.json | 38 +++ 13 files changed, 851 insertions(+), 47 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/broadwell/counter.json delete mode 100644 tools/perf/pmu-events/arch/x86/broadwell/uncore-other.json
tools/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json +43 -37
@@ -90,7 +90,7 @@ { "BriefDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists", "MetricExpr": "66 * OTHER_ASSISTS.ANY_WB_ASSIST / tma_info_thread_slots", - "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_group", + "MetricGroup": "BvIO;TopdownL4;tma_L4_group;tma_microcode_sequencer_group", "MetricName": "tma_assists", "MetricThreshold": "tma_assists > 0.1 & (tma_microcode_sequencer > 0.05 & tma_heavy_operations > 0.1)", "PublicDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists. Assists are long sequences of uops that are required in certain corner-cases for operations that cannot be handled natively by the execution pipeline. For example; when working with very small floating point values (so-called Denormals); the FP units are not set up to perform these operations natively. Instead; a sequence of instructions to perform the computation on the Denormals is injected into the pipeline. Since these microcode sequences might be dozens of uops long; Assists can be extremely deleterious to performance and they can be avoided in many cases. Sample with: OTHER_ASSISTS.ANY",
@@ -100,7 +100,7 @@ "BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend", "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma_retiring)", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", "MetricgroupNoGroup": "TopdownL1",
@@ -121,7 +121,7 @@ "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Branch Misprediction", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT) * tma_bad_speculation", - "MetricGroup": "BadSpec;BrMispredicts;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -160,7 +160,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(60 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) + 43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS)))) / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", "MetricThreshold": "tma_contested_accesses > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses. Contested accesses occur when data written by one Logical Processor are read by another Logical Processor on a different Physical Core. Examples of contested accesses include synchronizations such as locks; true data sharing such as modified locked variables; and false sharing. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM_PS;MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS_PS. Related metrics: tma_data_sharing, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -181,7 +181,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) / tma_info_thread_clks", - "MetricGroup": "Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_data_sharing", "MetricThreshold": "tma_data_sharing > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses. Data shared by multiple Logical Processors (even just read shared) may cause increased access latency due to cache coherency. Excessive data sharing can drastically harm multithreaded performance. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HIT_PS. Related metrics: tma_contested_accesses, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -190,7 +190,7 @@ { "BriefDescription": "This metric represents fraction of cycles where the Divider unit was active", "MetricExpr": "ARITH.FPU_DIV_ACTIVE / tma_info_core_core_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricGroup": "BvCB;TopdownL3;tma_L3_group;tma_core_bound_group", "MetricName": "tma_divider", "MetricThreshold": "tma_divider > 0.2 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2)", "PublicDescription": "This metric represents fraction of cycles where the Divider unit was active. Divide and square root instructions are performed by the Divider unit and can take considerably longer latency than integer or Floating Point addition; subtraction; or multiplication. Sample with: ARITH.DIVIDER_UOPS",
@@ -227,7 +227,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses", "MetricExpr": "(8 * DTLB_LOAD_MISSES.STLB_HIT + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_LOAD_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", "MetricName": "tma_dtlb_load", "MetricThreshold": "tma_dtlb_load > 0.1 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses. TLBs (Translation Look-aside Buffers) are processor caches for recently used entries out of the Page Tables that are used to map virtual- to physical-addresses by the operating system. This metric approximates the potential delay of demand loads missing the first-level data TLB (assuming worst case scenario with back to back misses to different pages). This includes hitting in the second-level TLB (STLB) as well as performing a hardware page walk on an STLB miss. Sample with: MEM_UOPS_RETIRED.STLB_MISS_LOADS_PS. Related metrics: tma_dtlb_store",
@@ -236,7 +236,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses", "MetricExpr": "(8 * DTLB_STORE_MISSES.STLB_HIT + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_STORE_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", "MetricName": "tma_dtlb_store", "MetricThreshold": "tma_dtlb_store > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses. As with ordinary data caching; focus on improving data locality and reducing working-set size to reduce DTLB overhead. Additionally; consider using profile-guided optimization (PGO) to collocate frequently-used data on the same page. Try using larger page sizes for large amounts of frequently-used data. Sample with: MEM_UOPS_RETIRED.STLB_MISS_STORES_PS. Related metrics: tma_dtlb_load",
@@ -245,7 +245,7 @@ { "BriefDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing", "MetricExpr": "60 * OFFCORE_RESPONSE.DEMAND_RFO.L3_HIT.SNOOP_HITM / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", "MetricName": "tma_false_sharing", "MetricThreshold": "tma_false_sharing > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing. False Sharing is a multithreading hiccup; where multiple Logical Processors contend on different data-elements mapped into the same cache line. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM_PS;OFFCORE_RESPONSE.DEMAND_RFO.L3_HIT.SNOOP_HITM. Related metrics: tma_contested_accesses, tma_data_sharing, tma_machine_clears, tma_remote_cache",
@@ -255,7 +255,7 @@ "BriefDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_info_memory_load_miss_real_latency * cpu@L1D_PEND_MISS.FB_FULL\\,cmask\\=1@ / tma_info_thread_clks", - "MetricGroup": "MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", + "MetricGroup": "BvMS;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", "PublicDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed. The higher the metric value; the deeper the memory hierarchy level the misses are satisfied from (metric values >1 are valid). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or external memory). Related metrics: tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_latency, tma_streaming_stores",
@@ -292,7 +292,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.SCALAR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -301,7 +301,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) vector uops fraction the CPU has retired aggregated across all vector widths", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.VECTOR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -329,7 +329,7 @@ { "BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend", "MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / tma_info_thread_slots", - "MetricGroup": "PGO;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvFB;BvIO;PGO;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", "MetricgroupNoGroup": "TopdownL1",
@@ -349,7 +349,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses.", "MetricExpr": "ICACHE.IFDATA_STALL / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_icache_misses", "MetricThreshold": "tma_icache_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "ScaleUnit": "100%"
@@ -388,7 +388,7 @@ }, { "BriefDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width)", - "MetricExpr": "(cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@) / (2 * tma_info_core_core_clks)", + "MetricExpr": "(FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR) / (2 * tma_info_core_core_clks)", "MetricGroup": "Cor;Flops;HPC", "MetricName": "tma_info_core_fp_arith_utilization", "PublicDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width). Values > 1 are possible due to ([BDW+] Fused-Multiply Add (FMA) counting - common; [ADL+] use all of ADD/MUL/FMA in Scalar or 128/256-bit vectors - less common)."
@@ -428,7 +428,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (lower number means higher occurrence rate)", - "MetricExpr": "INST_RETIRED.ANY / (cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@)", + "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10",
@@ -502,12 +502,12 @@ "MetricThreshold": "tma_info_inst_mix_ipstore < 8" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Instructions per taken branch", "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW;Frontend;PGO;tma_issueFB", "MetricName": "tma_info_inst_mix_iptb", "MetricThreshold": "tma_info_inst_mix_iptb < 9", - "PublicDescription": "Instruction per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" + "PublicDescription": "Instructions per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" }, { "BriefDescription": "Average per-core data fill bandwidth to the L1 data cache [GB / sec]",
@@ -528,7 +528,7 @@ "MetricName": "tma_info_memory_core_l3_cache_fill_bw_2t" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L1 data cache [GB / sec]", "MetricExpr": "64 * L1D.REPLACEMENT / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw"
@@ -540,7 +540,7 @@ "MetricName": "tma_info_memory_l1mpki" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L2 cache [GB / sec]", "MetricExpr": "64 * L2_LINES_IN.ALL / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l2_cache_fill_bw"
@@ -576,7 +576,13 @@ "MetricName": "tma_info_memory_l2mpki_load" }, { - "BriefDescription": "", + "BriefDescription": "Offcore requests (L2 cache miss) per kilo instruction for demand RFOs", + "MetricExpr": "1e3 * OFFCORE_REQUESTS.DEMAND_RFO / INST_RETIRED.ANY", + "MetricGroup": "CacheMisses;Offcore", + "MetricName": "tma_info_memory_l2mpki_rfo" + }, + { + "BriefDescription": "Average per-thread data fill bandwidth to the L3 cache [GB / sec]", "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l3_cache_fill_bw"
@@ -628,7 +634,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0.5" }, { - "BriefDescription": "", + "BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is execution) per core", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute"
@@ -647,13 +653,13 @@ }, { "BriefDescription": "Average CPU Utilization (percentage)", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "tma_info_system_cpus_utilized / #num_cpus_online", "MetricGroup": "HPC;Summary", "MetricName": "tma_info_system_cpu_utilization" }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "#num_cpus_online * tma_info_system_cpu_utilization", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" },
@@ -748,7 +754,7 @@ "MetricThreshold": "tma_info_thread_uoppi > 1.05" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Uops per taken branch", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW", "MetricName": "tma_info_thread_uptb",
@@ -757,7 +763,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses", "MetricExpr": "(14 * ITLB_MISSES.STLB_HIT + cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * ITLB_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_itlb_misses", "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses. Sample with: ITLB_MISSES.WALK_COMPLETED",
@@ -775,7 +781,7 @@ { "BriefDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks", - "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", + "MetricGroup": "BvML;CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", "MetricName": "tma_l2_bound", "MetricThreshold": "tma_l2_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2)", "PublicDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads. Avoiding cache misses (i.e. L1 misses/L2 hits) can improve the latency and increase performance. Sample with: MEM_LOAD_UOPS_RETIRED.L2_HIT_PS",
@@ -795,7 +801,7 @@ "BriefDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited)", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "29 * (MEM_LOAD_UOPS_RETIRED.L3_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", + "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3 hits) will improve the latency; reduce contention with sibling physical cores and increase performance. Note the value of this node may overlap with its siblings. Sample with: MEM_LOAD_UOPS_RETIRED.L3_HIT_PS. Related metrics: tma_mem_latency",
@@ -844,7 +850,7 @@ "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Machine Clears", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bad_speculation - tma_branch_mispredicts", - "MetricGroup": "BadSpec;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", + "MetricGroup": "BadSpec;BvMS;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", "MetricName": "tma_machine_clears", "MetricThreshold": "tma_machine_clears > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -854,7 +860,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, cpu@OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD\\,cmask\\=4@) / tma_info_thread_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuristic assumes that a similar off-core traffic is generated by all IA cores. This metric does not aggregate non-data-read requests by this logical processor; requests from other IA Logical Processors/Physical Cores/sockets; or other non-IA devices like GPU; hence the maximum external memory bandwidth limits may or may not be approached when this metric is flagged (see Uncore counters for that). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full",
@@ -863,7 +869,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD) / tma_info_thread_clks - tma_mem_bandwidth", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from other Logical Processors/Physical Cores/sockets (see Uncore counters for that). Related metrics: tma_l3_hit_latency",
@@ -892,7 +898,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES * tma_branch_resteers / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY)", - "MetricGroup": "BadSpec;BrMispredicts;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", "MetricThreshold": "tma_mispredicts_resteers > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage. Related metrics: tma_branch_mispredicts, tma_info_bad_spec_branch_misprediction_cost",
@@ -1028,7 +1034,7 @@ { "BriefDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise).", "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,cmask\\=3@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks", - "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", + "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", "MetricName": "tma_ports_utilized_3m", "MetricThreshold": "tma_ports_utilized_3m > 0.4 & (tma_ports_utilization > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", "ScaleUnit": "100%"
@@ -1036,7 +1042,7 @@ { "BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / tma_info_thread_slots", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvUW;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.1", "MetricgroupNoGroup": "TopdownL1",
@@ -1065,7 +1071,7 @@ { "BriefDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors)", "MetricExpr": "(OFFCORE_REQUESTS_BUFFER.SQ_FULL / 2 if #SMT_on else OFFCORE_REQUESTS_BUFFER.SQ_FULL) / tma_info_core_core_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_mem_bandwidth",
@@ -1093,7 +1099,7 @@ "BriefDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(L2_RQSTS.RFO_HIT * 9 * (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) + (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) * min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO)) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", "MetricName": "tma_store_latency", "MetricThreshold": "tma_store_latency > 0.1 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses. Store accesses usually less impact out-of-order core performance; however; holding resources for longer time can lead into undesired implications (e.g. contention on L1D fill-buffer entries - see FB_Full). Related metrics: tma_fb_full, tma_lock_latency",
@@ -1110,7 +1116,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears", "MetricExpr": "tma_branch_resteers - tma_mispredicts_resteers - tma_clears_resteers", - "MetricGroup": "BigFootprint;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", "MetricName": "tma_unknown_branches", "MetricThreshold": "tma_unknown_branches > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears. These are fetched branches the Branch Prediction Unit was unable to recognize (e.g. first time the branch is fetched or hitting BTB capacity limit) hence called Unknown Branches. Sample with: BACLEARS.ANY",
tools/perf/pmu-events/arch/x86/broadwell/cache.json +275 -0
1,925 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwell/counter.json new file +22 -0
@@ -0,0 +1,22 @@ +[ + { + "Unit": "core", + "CountersNumFixed": "3", + "CountersNumGeneric": "4" + }, + { + "Unit": "CBOX", + "CountersNumFixed": "0", + "CountersNumGeneric": "2" + }, + { + "Unit": "ARB", + "CountersNumFixed": "0", + "CountersNumGeneric": "2" + }, + { + "Unit": "cbox_0", + "CountersNumFixed": 1, + "CountersNumGeneric": "0" + } +]
tools/perf/pmu-events/arch/x86/broadwell/floating-point.json +22 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -17,6 +19,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -25,6 +28,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 8 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -33,6 +37,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single and 256-bit packed double precision FP instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, 1 for each element. Applies to SSE* and AVX* packed single precision and packed double precision FP instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision and 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -41,6 +46,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.DOUBLE", "SampleAfterValue": "2000006",
@@ -48,6 +54,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* packed double and single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.PACKED", "SampleAfterValue": "2000004",
@@ -55,6 +62,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computation operation. Applies to SSE* and AVX* scalar double and single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", "PublicDescription": "Number of SSE/AVX computational scalar single precision and double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -63,6 +71,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", "PublicDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -71,6 +80,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", "PublicDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -79,6 +89,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SINGLE", "SampleAfterValue": "2000005",
@@ -86,6 +97,7 @@ }, { "BriefDescription": "Number of any Vector retired FP arithmetic instructions", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", "SampleAfterValue": "2000003",
@@ -93,6 +105,7 @@ }, { "BriefDescription": "Cycles with any input/output SSE or FP assist", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0xCA", "EventName": "FP_ASSIST.ANY",
@@ -102,6 +115,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to input values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_INPUT", "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
@@ -110,6 +124,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to Output values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_OUTPUT", "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
@@ -118,6 +133,7 @@ }, { "BriefDescription": "Number of X87 assists due to input value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_INPUT", "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
@@ -126,6 +142,7 @@ }, { "BriefDescription": "Number of X87 assists due to output value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_OUTPUT", "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
@@ -134,6 +151,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_ELIMINATED", "SampleAfterValue": "1000003",
@@ -141,6 +159,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were not eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_NOT_ELIMINATED", "SampleAfterValue": "1000003",
@@ -148,6 +167,7 @@ }, { "BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.AVX_TO_SSE",
@@ -157,6 +177,7 @@ }, { "BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.SSE_TO_AVX",
@@ -166,6 +187,7 @@ }, { "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports", + "Counter": "0,1,2,3", "EventCode": "0xA0", "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF", "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
tools/perf/pmu-events/arch/x86/broadwell/frontend.json +28 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.", + "Counter": "0,1,2,3", "EventCode": "0xe6", "EventName": "BACLEARS.ANY", "SampleAfterValue": "100003",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.", + "Counter": "0,1,2,3", "EventCode": "0xAB", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
@@ -16,6 +18,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Reads. both cacheable and noncacheable, including UC fetches", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.HIT", "PublicDescription": "This event counts the number of both cacheable and noncacheable Instruction Cache, Streaming Buffer and Victim Cache Reads including UC fetches.",
@@ -24,6 +27,7 @@ }, { "BriefDescription": "Cycles where a code fetch is stalled due to L1 instruction-cache miss.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.IFDATA_STALL", "PublicDescription": "This event counts cycles during which the demand fetch waits for data (wfdM104H) from L2 or iSB (opportunistic hit).",
@@ -32,6 +36,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Misses. Includes Uncacheable accesses.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.MISSES", "PublicDescription": "This event counts the number of instruction cache, streaming buffer and victim cache misses. Counting includes UC accesses.",
@@ -40,6 +45,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
@@ -49,6 +55,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
@@ -58,6 +65,7 @@ }, { "BriefDescription": "Cycles MITE is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
@@ -67,6 +75,7 @@ }, { "BriefDescription": "Cycles MITE is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
@@ -76,6 +85,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES",
@@ -85,6 +95,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
@@ -93,6 +104,7 @@ }, { "BriefDescription": "Instruction Decode Queue (IDQ) empty cycles", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.EMPTY", "PublicDescription": "This counts the number of cycles that the instruction decoder queue is empty and can indicate that the application may be bound in the front end. It does not determine whether there are uops being delivered to the Alloc stage since uops can be delivered by bypass skipping the Instruction Decode Queue (IDQ) when it is empty.",
@@ -101,6 +113,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_ALL_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -109,6 +122,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES",
@@ -118,6 +132,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -126,6 +141,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES",
@@ -135,6 +151,7 @@ }, { "BriefDescription": "Cycles when uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_CYCLES",
@@ -144,6 +161,7 @@ }, { "BriefDescription": "Deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -154,6 +172,7 @@ }, { "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_UOPS", "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -162,6 +181,7 @@ }, { "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_MITE_UOPS", "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -170,6 +190,7 @@ }, { "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -179,6 +200,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
@@ -187,6 +209,7 @@ }, { "BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled", + "Counter": "0,1,2,3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
@@ -195,6 +218,7 @@ }, { "BriefDescription": "Cycles per thread when 4 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE",
@@ -204,6 +228,7 @@ }, { "BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
@@ -213,6 +238,7 @@ }, { "BriefDescription": "Cycles per thread when 3 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_1_UOP_DELIV.CORE",
@@ -222,6 +248,7 @@ }, { "BriefDescription": "Cycles with less than 2 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "2", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_2_UOP_DELIV.CORE",
@@ -230,6 +257,7 @@ }, { "BriefDescription": "Cycles with less than 3 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
tools/perf/pmu-events/arch/x86/broadwell/memory.json +240 -0
403 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwell/metricgroups.json +11 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/broadwell/other.json +4 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/broadwell/pipeline.json +137 -0
137 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwell/uncore-cache.json +24 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/broadwell/uncore-interconnect.json +7 -0
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/broadwell/uncore-other.json deleted +0 -10
Lines of this file are not shown. View the raw message.
tools/perf/pmu-events/arch/x86/broadwell/virtual-memory.json +38 -0
Lines of this file are not shown. View the raw message.
The rest of this diff is not shown. View the raw message.
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Part of the diff is not shown here. View the raw message.
Ian Rogers
reviewerPATCH 5/37 perf vendor events: Update broadwellde metrics add event counter information
In reply to Ian RogersAdd counter information necessary for optimizing event grouping the perf tool.
The most recent RFC patch set using this information:
https://lore.kernel.org/lkml/[email protected]/
The information was added in:
https://github.com/intel/perfmon/commit/475892a9690cb048949e593fe39cee65cd4765e1
and later patches.
The TMA 4.8 information was updated in:
https://github.com/intel/perfmon/commit/59194d4d90ca50a3fcb2de0d82b9f6fc0c9a5736
--- .../arch/x86/broadwellde/bdwde-metrics.json | 80 ++-- .../arch/x86/broadwellde/cache.json | 76 ++++ .../arch/x86/broadwellde/counter.json | 42 ++ .../arch/x86/broadwellde/floating-point.json | 22 + .../arch/x86/broadwellde/frontend.json | 28 ++ .../arch/x86/broadwellde/memory.json | 39 ++ .../arch/x86/broadwellde/metricgroups.json | 11 + .../arch/x86/broadwellde/other.json | 4 + .../arch/x86/broadwellde/pipeline.json | 137 +++++++ .../arch/x86/broadwellde/uncore-cache.json | 382 ++++++++++++++++++ .../x86/broadwellde/uncore-interconnect.json | 70 ++++ .../arch/x86/broadwellde/uncore-io.json | 62 +++ .../arch/x86/broadwellde/uncore-memory.json | 322 +++++++++++++++ .../arch/x86/broadwellde/uncore-power.json | 57 +++ .../arch/x86/broadwellde/virtual-memory.json | 38 ++ 15 files changed, 1333 insertions(+), 37 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/broadwellde/counter.json
tools/perf/pmu-events/arch/x86/broadwellde/bdwde-metrics.json +43 -37
@@ -90,7 +90,7 @@ { "BriefDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists", "MetricExpr": "66 * OTHER_ASSISTS.ANY_WB_ASSIST / tma_info_thread_slots", - "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_group", + "MetricGroup": "BvIO;TopdownL4;tma_L4_group;tma_microcode_sequencer_group", "MetricName": "tma_assists", "MetricThreshold": "tma_assists > 0.1 & (tma_microcode_sequencer > 0.05 & tma_heavy_operations > 0.1)", "PublicDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists. Assists are long sequences of uops that are required in certain corner-cases for operations that cannot be handled natively by the execution pipeline. For example; when working with very small floating point values (so-called Denormals); the FP units are not set up to perform these operations natively. Instead; a sequence of instructions to perform the computation on the Denormals is injected into the pipeline. Since these microcode sequences might be dozens of uops long; Assists can be extremely deleterious to performance and they can be avoided in many cases. Sample with: ASSISTS.ANY",
@@ -100,7 +100,7 @@ "BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend", "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma_retiring)", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", "MetricgroupNoGroup": "TopdownL1",
@@ -121,7 +121,7 @@ "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Branch Misprediction", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT) * tma_bad_speculation", - "MetricGroup": "BadSpec;BrMispredicts;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -160,7 +160,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(60 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) + 43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS)))) / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", "MetricThreshold": "tma_contested_accesses > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses. Contested accesses occur when data written by one Logical Processor are read by another Logical Processor on a different Physical Core. Examples of contested accesses include synchronizations such as locks; true data sharing such as modified locked variables; and false sharing. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD;MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS. Related metrics: tma_data_sharing, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -181,7 +181,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) / tma_info_thread_clks", - "MetricGroup": "Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_data_sharing", "MetricThreshold": "tma_data_sharing > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses. Data shared by multiple Logical Processors (even just read shared) may cause increased access latency due to cache coherency. Excessive data sharing can drastically harm multithreaded performance. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_NO_FWD. Related metrics: tma_contested_accesses, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -190,7 +190,7 @@ { "BriefDescription": "This metric represents fraction of cycles where the Divider unit was active", "MetricExpr": "ARITH.FPU_DIV_ACTIVE / tma_info_core_core_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricGroup": "BvCB;TopdownL3;tma_L3_group;tma_core_bound_group", "MetricName": "tma_divider", "MetricThreshold": "tma_divider > 0.2 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2)", "PublicDescription": "This metric represents fraction of cycles where the Divider unit was active. Divide and square root instructions are performed by the Divider unit and can take considerably longer latency than integer or Floating Point addition; subtraction; or multiplication. Sample with: ARITH.DIVIDER_ACTIVE",
@@ -227,7 +227,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses", "MetricExpr": "(8 * DTLB_LOAD_MISSES.STLB_HIT + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_LOAD_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", "MetricName": "tma_dtlb_load", "MetricThreshold": "tma_dtlb_load > 0.1 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses. TLBs (Translation Look-aside Buffers) are processor caches for recently used entries out of the Page Tables that are used to map virtual- to physical-addresses by the operating system. This metric approximates the potential delay of demand loads missing the first-level data TLB (assuming worst case scenario with back to back misses to different pages). This includes hitting in the second-level TLB (STLB) as well as performing a hardware page walk on an STLB miss. Sample with: MEM_INST_RETIRED.STLB_MISS_LOADS_PS. Related metrics: tma_dtlb_store",
@@ -236,7 +236,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses", "MetricExpr": "(8 * DTLB_STORE_MISSES.STLB_HIT + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_STORE_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", "MetricName": "tma_dtlb_store", "MetricThreshold": "tma_dtlb_store > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses. As with ordinary data caching; focus on improving data locality and reducing working-set size to reduce DTLB overhead. Additionally; consider using profile-guided optimization (PGO) to collocate frequently-used data on the same page. Try using larger page sizes for large amounts of frequently-used data. Sample with: MEM_INST_RETIRED.STLB_MISS_STORES_PS. Related metrics: tma_dtlb_load",
@@ -246,7 +246,7 @@ "BriefDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_info_memory_load_miss_real_latency * cpu@L1D_PEND_MISS.FB_FULL\\,cmask\\=1@ / tma_info_thread_clks", - "MetricGroup": "MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", + "MetricGroup": "BvMS;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", "PublicDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed. The higher the metric value; the deeper the memory hierarchy level the misses are satisfied from (metric values >1 are valid). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or external memory). Related metrics: tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_latency, tma_streaming_stores",
@@ -283,7 +283,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.SCALAR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -292,7 +292,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) vector uops fraction the CPU has retired aggregated across all vector widths", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.VECTOR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -320,7 +320,7 @@ { "BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend", "MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / tma_info_thread_slots", - "MetricGroup": "PGO;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvFB;BvIO;PGO;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", "MetricgroupNoGroup": "TopdownL1",
@@ -340,7 +340,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses", "MetricExpr": "ICACHE.IFDATA_STALL / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_icache_misses", "MetricThreshold": "tma_icache_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses. Sample with: FRONTEND_RETIRED.L2_MISS_PS;FRONTEND_RETIRED.L1I_MISS_PS",
@@ -380,7 +380,7 @@ }, { "BriefDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width)", - "MetricExpr": "(cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@) / (2 * tma_info_core_core_clks)", + "MetricExpr": "(FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR) / (2 * tma_info_core_core_clks)", "MetricGroup": "Cor;Flops;HPC", "MetricName": "tma_info_core_fp_arith_utilization", "PublicDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width). Values > 1 are possible due to ([BDW+] Fused-Multiply Add (FMA) counting - common; [ADL+] use all of ADD/MUL/FMA in Scalar or 128/256-bit vectors - less common)."
@@ -420,7 +420,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (lower number means higher occurrence rate)", - "MetricExpr": "INST_RETIRED.ANY / (cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@)", + "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10",
@@ -494,12 +494,12 @@ "MetricThreshold": "tma_info_inst_mix_ipstore < 8" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Instructions per taken branch", "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW;Frontend;PGO;tma_issueFB", "MetricName": "tma_info_inst_mix_iptb", "MetricThreshold": "tma_info_inst_mix_iptb < 9", - "PublicDescription": "Instruction per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" + "PublicDescription": "Instructions per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" }, { "BriefDescription": "Average per-core data fill bandwidth to the L1 data cache [GB / sec]",
@@ -520,7 +520,7 @@ "MetricName": "tma_info_memory_core_l3_cache_fill_bw_2t" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L1 data cache [GB / sec]", "MetricExpr": "64 * L1D.REPLACEMENT / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw"
@@ -532,7 +532,7 @@ "MetricName": "tma_info_memory_l1mpki" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L2 cache [GB / sec]", "MetricExpr": "64 * L2_LINES_IN.ALL / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l2_cache_fill_bw"
@@ -568,7 +568,13 @@ "MetricName": "tma_info_memory_l2mpki_load" }, { - "BriefDescription": "", + "BriefDescription": "Offcore requests (L2 cache miss) per kilo instruction for demand RFOs", + "MetricExpr": "1e3 * OFFCORE_REQUESTS.DEMAND_RFO / INST_RETIRED.ANY", + "MetricGroup": "CacheMisses;Offcore", + "MetricName": "tma_info_memory_l2mpki_rfo" + }, + { + "BriefDescription": "Average per-thread data fill bandwidth to the L3 cache [GB / sec]", "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l3_cache_fill_bw"
@@ -620,7 +626,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0.5" }, { - "BriefDescription": "", + "BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is execution) per core", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute"
@@ -639,13 +645,13 @@ }, { "BriefDescription": "Average CPU Utilization (percentage)", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "tma_info_system_cpus_utilized / #num_cpus_online", "MetricGroup": "HPC;Summary", "MetricName": "tma_info_system_cpu_utilization" }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "#num_cpus_online * tma_info_system_cpu_utilization", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" },
@@ -740,7 +746,7 @@ "MetricThreshold": "tma_info_thread_uoppi > 1.05" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Uops per taken branch", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW", "MetricName": "tma_info_thread_uptb",
@@ -749,7 +755,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses", "MetricExpr": "(14 * ITLB_MISSES.STLB_HIT + cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * ITLB_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_itlb_misses", "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses. Sample with: FRONTEND_RETIRED.STLB_MISS_PS;FRONTEND_RETIRED.ITLB_MISS_PS",
@@ -767,7 +773,7 @@ { "BriefDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks", - "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", + "MetricGroup": "BvML;CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", "MetricName": "tma_l2_bound", "MetricThreshold": "tma_l2_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2)", "PublicDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads. Avoiding cache misses (i.e. L1 misses/L2 hits) can improve the latency and increase performance. Sample with: MEM_LOAD_RETIRED.L2_HIT_PS",
@@ -787,7 +793,7 @@ "BriefDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited)", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "29 * (MEM_LOAD_UOPS_RETIRED.L3_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_RETIRED.L3_MISS))) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", + "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3 hits) will improve the latency; reduce contention with sibling physical cores and increase performance. Note the value of this node may overlap with its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tma_mem_latency",
@@ -829,14 +835,14 @@ "MetricGroup": "Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_l1_bound_group", "MetricName": "tma_lock_latency", "MetricThreshold": "tma_lock_latency > 0.2 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric represents fraction of cycles the CPU spent handling cache misses due to lock operations. Due to the microarchitecture handling of locks; they are classified as L1_Bound regardless of what memory source satisfied them. Sample with: MEM_INST_RETIRED.LOCK_LOADS_PS. Related metrics: tma_store_latency", + "PublicDescription": "This metric represents fraction of cycles the CPU spent handling cache misses due to lock operations. Due to the microarchitecture handling of locks; they are classified as L1_Bound regardless of what memory source satisfied them. Sample with: MEM_INST_RETIRED.LOCK_LOADS. Related metrics: tma_store_latency", "ScaleUnit": "100%" }, { "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Machine Clears", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bad_speculation - tma_branch_mispredicts", - "MetricGroup": "BadSpec;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", + "MetricGroup": "BadSpec;BvMS;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", "MetricName": "tma_machine_clears", "MetricThreshold": "tma_machine_clears > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -846,7 +852,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, cpu@OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD\\,cmask\\=4@) / tma_info_thread_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuristic assumes that a similar off-core traffic is generated by all IA cores. This metric does not aggregate non-data-read requests by this logical processor; requests from other IA Logical Processors/Physical Cores/sockets; or other non-IA devices like GPU; hence the maximum external memory bandwidth limits may or may not be approached when this metric is flagged (see Uncore counters for that). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full",
@@ -855,7 +861,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD) / tma_info_thread_clks - tma_mem_bandwidth", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from other Logical Processors/Physical Cores/sockets (see Uncore counters for that). Related metrics: tma_l3_hit_latency",
@@ -884,7 +890,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES * tma_branch_resteers / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY)", - "MetricGroup": "BadSpec;BrMispredicts;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", "MetricThreshold": "tma_mispredicts_resteers > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage. Sample with: INT_MISC.CLEAR_RESTEER_CYCLES. Related metrics: tma_branch_mispredicts, tma_info_bad_spec_branch_misprediction_cost",
@@ -1017,7 +1023,7 @@ { "BriefDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise)", "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,cmask\\=3@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks", - "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", + "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", "MetricName": "tma_ports_utilized_3m", "MetricThreshold": "tma_ports_utilized_3m > 0.4 & (tma_ports_utilization > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", "PublicDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise). Sample with: UOPS_EXECUTED.CYCLES_GE_3",
@@ -1026,7 +1032,7 @@ { "BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / tma_info_thread_slots", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvUW;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.1", "MetricgroupNoGroup": "TopdownL1",
@@ -1055,7 +1061,7 @@ { "BriefDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors)", "MetricExpr": "(OFFCORE_REQUESTS_BUFFER.SQ_FULL / 2 if #SMT_on else OFFCORE_REQUESTS_BUFFER.SQ_FULL) / tma_info_core_core_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_mem_bandwidth",
@@ -1083,7 +1089,7 @@ "BriefDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(L2_RQSTS.RFO_HIT * 9 * (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) + (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) * min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO)) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", "MetricName": "tma_store_latency", "MetricThreshold": "tma_store_latency > 0.1 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses. Store accesses usually less impact out-of-order core performance; however; holding resources for longer time can lead into undesired implications (e.g. contention on L1D fill-buffer entries - see FB_Full). Related metrics: tma_fb_full, tma_lock_latency",
@@ -1101,7 +1107,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears", "MetricExpr": "tma_branch_resteers - tma_mispredicts_resteers - tma_clears_resteers", - "MetricGroup": "BigFootprint;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", "MetricName": "tma_unknown_branches", "MetricThreshold": "tma_unknown_branches > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears. These are fetched branches the Branch Prediction Unit was unable to recognize (e.g. first time the branch is fetched or hitting BTB capacity limit) hence called Unknown Branches. Sample with: FRONTEND_RETIRED.UNKNOWN_BRANCH",
tools/perf/pmu-events/arch/x86/broadwellde/cache.json +76 -0
532 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/counter.json new file +42 -0
43 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/floating-point.json +22 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -17,6 +19,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -25,6 +28,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 8 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -33,6 +37,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single and 256-bit packed double precision FP instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, 1 for each element. Applies to SSE* and AVX* packed single precision and packed double precision FP instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision and 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -41,6 +46,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.DOUBLE", "SampleAfterValue": "2000006",
@@ -48,6 +54,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* packed double and single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.PACKED", "SampleAfterValue": "2000004",
@@ -55,6 +62,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computation operation. Applies to SSE* and AVX* scalar double and single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", "PublicDescription": "Number of SSE/AVX computational scalar single precision and double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -63,6 +71,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", "PublicDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -71,6 +80,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", "PublicDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -79,6 +89,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SINGLE", "SampleAfterValue": "2000005",
@@ -86,6 +97,7 @@ }, { "BriefDescription": "Number of any Vector retired FP arithmetic instructions", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", "SampleAfterValue": "2000003",
@@ -93,6 +105,7 @@ }, { "BriefDescription": "Cycles with any input/output SSE or FP assist", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0xCA", "EventName": "FP_ASSIST.ANY",
@@ -102,6 +115,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to input values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_INPUT", "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
@@ -110,6 +124,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to Output values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_OUTPUT", "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
@@ -118,6 +133,7 @@ }, { "BriefDescription": "Number of X87 assists due to input value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_INPUT", "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
@@ -126,6 +142,7 @@ }, { "BriefDescription": "Number of X87 assists due to output value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_OUTPUT", "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
@@ -134,6 +151,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_ELIMINATED", "SampleAfterValue": "1000003",
@@ -141,6 +159,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were not eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_NOT_ELIMINATED", "SampleAfterValue": "1000003",
@@ -148,6 +167,7 @@ }, { "BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.AVX_TO_SSE",
@@ -157,6 +177,7 @@ }, { "BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.SSE_TO_AVX",
@@ -166,6 +187,7 @@ }, { "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports", + "Counter": "0,1,2,3", "EventCode": "0xA0", "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF", "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
tools/perf/pmu-events/arch/x86/broadwellde/frontend.json +28 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.", + "Counter": "0,1,2,3", "EventCode": "0xe6", "EventName": "BACLEARS.ANY", "SampleAfterValue": "100003",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.", + "Counter": "0,1,2,3", "EventCode": "0xAB", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
@@ -16,6 +18,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Reads. both cacheable and noncacheable, including UC fetches", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.HIT", "PublicDescription": "This event counts the number of both cacheable and noncacheable Instruction Cache, Streaming Buffer and Victim Cache Reads including UC fetches.",
@@ -24,6 +27,7 @@ }, { "BriefDescription": "Cycles where a code fetch is stalled due to L1 instruction-cache miss.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.IFDATA_STALL", "PublicDescription": "This event counts cycles during which the demand fetch waits for data (wfdM104H) from L2 or iSB (opportunistic hit).",
@@ -32,6 +36,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Misses. Includes Uncacheable accesses.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.MISSES", "PublicDescription": "This event counts the number of instruction cache, streaming buffer and victim cache misses. Counting includes UC accesses.",
@@ -40,6 +45,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
@@ -49,6 +55,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
@@ -58,6 +65,7 @@ }, { "BriefDescription": "Cycles MITE is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
@@ -67,6 +75,7 @@ }, { "BriefDescription": "Cycles MITE is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
@@ -76,6 +85,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES",
@@ -85,6 +95,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
@@ -93,6 +104,7 @@ }, { "BriefDescription": "Instruction Decode Queue (IDQ) empty cycles", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.EMPTY", "PublicDescription": "This counts the number of cycles that the instruction decoder queue is empty and can indicate that the application may be bound in the front end. It does not determine whether there are uops being delivered to the Alloc stage since uops can be delivered by bypass skipping the Instruction Decode Queue (IDQ) when it is empty.",
@@ -101,6 +113,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_ALL_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -109,6 +122,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES",
@@ -118,6 +132,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -126,6 +141,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES",
@@ -135,6 +151,7 @@ }, { "BriefDescription": "Cycles when uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_CYCLES",
@@ -144,6 +161,7 @@ }, { "BriefDescription": "Deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -154,6 +172,7 @@ }, { "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_UOPS", "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -162,6 +181,7 @@ }, { "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_MITE_UOPS", "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -170,6 +190,7 @@ }, { "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -179,6 +200,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
@@ -187,6 +209,7 @@ }, { "BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled", + "Counter": "0,1,2,3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
@@ -195,6 +218,7 @@ }, { "BriefDescription": "Cycles per thread when 4 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE",
@@ -204,6 +228,7 @@ }, { "BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
@@ -213,6 +238,7 @@ }, { "BriefDescription": "Cycles per thread when 3 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_1_UOP_DELIV.CORE",
@@ -222,6 +248,7 @@ }, { "BriefDescription": "Cycles with less than 2 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "2", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_2_UOP_DELIV.CORE",
@@ -230,6 +257,7 @@ }, { "BriefDescription": "Cycles with less than 3 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
tools/perf/pmu-events/arch/x86/broadwellde/memory.json +39 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Number of times HLE abort was triggered", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED", "PEBS": "1",
@@ -10,6 +11,7 @@ }, { "BriefDescription": "Number of times an HLE execution aborted due to various memory events (e.g., read/write capacity and conflicts).", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED_MISC1", "PublicDescription": "Number of times an HLE abort was attributed to a Memory condition (See TSX_Memory event for additional details).",
@@ -18,6 +20,7 @@ }, { "BriefDescription": "Number of times an HLE execution aborted due to uncommon conditions", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED_MISC2", "PublicDescription": "Number of times the TSX watchdog signaled an HLE abort.",
@@ -26,6 +29,7 @@ }, { "BriefDescription": "Number of times an HLE execution aborted due to HLE-unfriendly instructions", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED_MISC3", "PublicDescription": "Number of times a disallowed operation caused an HLE abort.",
@@ -34,6 +38,7 @@ }, { "BriefDescription": "Number of times an HLE execution aborted due to incompatible memory type", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED_MISC4", "PublicDescription": "Number of times HLE caused a fault.",
@@ -42,6 +47,7 @@ }, { "BriefDescription": "Number of times an HLE execution aborted due to none of the previous 4 categories (e.g. interrupts)", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.ABORTED_MISC5", "PublicDescription": "Number of times HLE aborted and was not due to the abort conditions in subevents 3-6.",
@@ -50,6 +56,7 @@ }, { "BriefDescription": "Number of times HLE commit succeeded", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.COMMIT", "PublicDescription": "Number of times HLE commit succeeded.",
@@ -58,6 +65,7 @@ }, { "BriefDescription": "Number of times we entered an HLE region; does not count nested transactions", + "Counter": "0,1,2,3", "EventCode": "0xc8", "EventName": "HLE_RETIRED.START", "PublicDescription": "Number of times we entered an HLE region\n does not count nested transactions.",
@@ -66,6 +74,7 @@ }, { "BriefDescription": "Counts the number of machine clears due to memory order conflicts.", + "Counter": "0,1,2,3", "EventCode": "0xC3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", "PublicDescription": "This event counts the number of memory ordering Machine Clears detected. Memory Ordering Machine Clears can result from one of the following:\n1. memory disambiguation,\n2. external snoop, or\n3. cross SMT-HW-thread snoop (stores) hitting load buffer.",
@@ -74,6 +83,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 128", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -87,6 +97,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 16", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -100,6 +111,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 256", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -113,6 +125,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 32", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -126,6 +139,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 4", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -139,6 +153,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 512", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -152,6 +167,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 64", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -165,6 +181,7 @@ }, { "BriefDescription": "Randomly selected loads with latency value being above 8", + "Counter": "3", "Data_LA": "1", "Errata": "BDM100, BDM35", "EventCode": "0xcd",
@@ -178,6 +195,7 @@ }, { "BriefDescription": "Speculative cache line split load uops dispatched to L1 cache", + "Counter": "0,1,2,3", "EventCode": "0x05", "EventName": "MISALIGN_MEM_REF.LOADS", "PublicDescription": "This event counts speculative cache-line split load uops dispatched to the L1 cache.",
@@ -186,6 +204,7 @@ }, { "BriefDescription": "Speculative cache line split STA uops dispatched to L1 cache", + "Counter": "0,1,2,3", "EventCode": "0x05", "EventName": "MISALIGN_MEM_REF.STORES", "PublicDescription": "This event counts speculative cache line split store-address (STA) uops dispatched to the L1 cache.",
@@ -194,6 +213,7 @@ }, { "BriefDescription": "Number of times RTM abort was triggered", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED", "PEBS": "1",
@@ -203,6 +223,7 @@ }, { "BriefDescription": "Number of times an RTM execution aborted due to various memory events (e.g. read/write capacity and conflicts)", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MISC1", "PublicDescription": "Number of times an RTM abort was attributed to a Memory condition (See TSX_Memory event for additional details).",
@@ -211,6 +232,7 @@ }, { "BriefDescription": "Number of times an RTM execution aborted due to various memory events (e.g., read/write capacity and conflicts).", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MISC2", "PublicDescription": "Number of times the TSX watchdog signaled an RTM abort.",
@@ -219,6 +241,7 @@ }, { "BriefDescription": "Number of times an RTM execution aborted due to HLE-unfriendly instructions", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MISC3", "PublicDescription": "Number of times a disallowed operation caused an RTM abort.",
@@ -227,6 +250,7 @@ }, { "BriefDescription": "Number of times an RTM execution aborted due to incompatible memory type", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MISC4", "PublicDescription": "Number of times a RTM caused a fault.",
@@ -235,6 +259,7 @@ }, { "BriefDescription": "Number of times an RTM execution aborted due to none of the previous 4 categories (e.g. interrupt)", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MISC5", "PublicDescription": "Number of times RTM aborted and was not due to the abort conditions in subevents 3-6.",
@@ -243,6 +268,7 @@ }, { "BriefDescription": "Number of times RTM commit succeeded", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.COMMIT", "PublicDescription": "Number of times RTM commit succeeded.",
@@ -251,6 +277,7 @@ }, { "BriefDescription": "Number of times we entered an RTM region; does not count nested transactions", + "Counter": "0,1,2,3", "EventCode": "0xc9", "EventName": "RTM_RETIRED.START", "PublicDescription": "Number of times we entered an RTM region\n does not count nested transactions.",
@@ -259,6 +286,7 @@ }, { "BriefDescription": "Counts the number of times a class of instructions that may cause a transactional abort was executed. Since this is the count of execution, it may not always cause a transactional abort.", + "Counter": "0,1,2,3", "EventCode": "0x5d", "EventName": "TX_EXEC.MISC1", "SampleAfterValue": "2000003",
@@ -266,6 +294,7 @@ }, { "BriefDescription": "Counts the number of times a class of instructions (e.g., vzeroupper) that may cause a transactional abort was executed inside a transactional region", + "Counter": "0,1,2,3", "EventCode": "0x5d", "EventName": "TX_EXEC.MISC2", "PublicDescription": "Unfriendly TSX abort triggered by a vzeroupper instruction.",
@@ -274,6 +303,7 @@ }, { "BriefDescription": "Counts the number of times an instruction execution caused the transactional nest count supported to be exceeded", + "Counter": "0,1,2,3", "EventCode": "0x5d", "EventName": "TX_EXEC.MISC3", "PublicDescription": "Unfriendly TSX abort triggered by a nest count that is too deep.",
@@ -282,6 +312,7 @@ }, { "BriefDescription": "Counts the number of times a XBEGIN instruction was executed inside an HLE transactional region.", + "Counter": "0,1,2,3", "EventCode": "0x5d", "EventName": "TX_EXEC.MISC4", "PublicDescription": "RTM region detected inside HLE.",
@@ -290,6 +321,7 @@ }, { "BriefDescription": "Counts the number of times an HLE XACQUIRE instruction was executed inside an RTM transactional region.", + "Counter": "0,1,2,3", "EventCode": "0x5d", "EventName": "TX_EXEC.MISC5", "SampleAfterValue": "2000003",
@@ -297,6 +329,7 @@ }, { "BriefDescription": "Number of times a TSX Abort was triggered due to an evicted line caused by a transaction overflow", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_WRITE", "PublicDescription": "Number of times a TSX Abort was triggered due to an evicted line caused by a transaction overflow.",
@@ -305,6 +338,7 @@ }, { "BriefDescription": "Number of times a TSX line had a cache conflict", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CONFLICT", "PublicDescription": "Number of times a TSX line had a cache conflict.",
@@ -313,6 +347,7 @@ }, { "BriefDescription": "Number of times a TSX Abort was triggered due to release/commit but data and address mismatch", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_HLE_ELISION_BUFFER_MISMATCH", "PublicDescription": "Number of times a TSX Abort was triggered due to release/commit but data and address mismatch.",
@@ -321,6 +356,7 @@ }, { "BriefDescription": "Number of times a TSX Abort was triggered due to commit but Lock Buffer not empty", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_HLE_ELISION_BUFFER_NOT_EMPTY", "PublicDescription": "Number of times a TSX Abort was triggered due to commit but Lock Buffer not empty.",
@@ -329,6 +365,7 @@ }, { "BriefDescription": "Number of times a TSX Abort was triggered due to attempting an unsupported alignment from Lock Buffer", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_HLE_ELISION_BUFFER_UNSUPPORTED_ALIGNMENT", "PublicDescription": "Number of times a TSX Abort was triggered due to attempting an unsupported alignment from Lock Buffer.",
@@ -337,6 +374,7 @@ }, { "BriefDescription": "Number of times a TSX Abort was triggered due to a non-release/commit store to lock", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_HLE_STORE_TO_ELIDED_LOCK", "PublicDescription": "Number of times a TSX Abort was triggered due to a non-release/commit store to lock.",
@@ -345,6 +383,7 @@ }, { "BriefDescription": "Number of times we could not allocate Lock Buffer", + "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.HLE_ELISION_BUFFER_FULL", "PublicDescription": "Number of times we could not allocate Lock Buffer.",
tools/perf/pmu-events/arch/x86/broadwellde/metricgroups.json +11 -0
@@ -5,7 +5,18 @@ "BigFootprint": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "BrMispredicts": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Branches": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvBC": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvCB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvFB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvIO": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvML": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMP": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMS": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMT": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvOB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvUW": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "CacheHits": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "CacheMisses": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Compute": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Cor": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "DSB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet",
tools/perf/pmu-events/arch/x86/broadwellde/other.json +4 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Unhalted core cycles when the thread is in ring 0", + "Counter": "0,1,2,3", "EventCode": "0x5C", "EventName": "CPL_CYCLES.RING0", "PublicDescription": "This event counts the unhalted core cycles during which the thread is in the ring 0 privileged mode.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Number of intervals between processor halts while thread is in ring 0", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x5C",
@@ -19,6 +21,7 @@ }, { "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3", + "Counter": "0,1,2,3", "EventCode": "0x5C", "EventName": "CPL_CYCLES.RING123", "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
@@ -27,6 +30,7 @@ }, { "BriefDescription": "Cycles when L1 and L2 are locked due to UC or split lock", + "Counter": "0,1,2,3", "EventCode": "0x63", "EventName": "LOCK_CYCLES.SPLIT_LOCK_UC_LOCK_DURATION", "PublicDescription": "This event counts cycles in which the L1 and L2 are locked due to a UC lock or split lock. A lock is asserted in case of locked memory access, due to noncacheable memory, locked operation that spans two cache lines, or a page walk from the noncacheable page table. L1D and L2 locks have a very high performance penalty and it is highly recommended to avoid such access.",
tools/perf/pmu-events/arch/x86/broadwellde/pipeline.json +137 -0
959 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/uncore-cache.json +382 -0
498 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/uncore-interconnect.json +70 -0
70 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/uncore-io.json +62 -0
62 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/uncore-memory.json +322 -0
322 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json +57 -0
57 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellde/virtual-memory.json +38 -0
Lines of this file are not shown. View the raw message.
The rest of this diff is not shown. View the raw message.
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Part of the diff is not shown here. View the raw message.
Ian Rogers
reviewerPATCH 6/37 perf vendor events: Update broadwellx metrics add event counter information
In reply to Ian RogersAdd counter information necessary for optimizing event grouping the perf tool.
The most recent RFC patch set using this information:
https://lore.kernel.org/lkml/[email protected]/
The information was added in:
https://github.com/intel/perfmon/commit/475892a9690cb048949e593fe39cee65cd4765e1
and later patches.
The TMA 4.8 information was updated in:
https://github.com/intel/perfmon/commit/59194d4d90ca50a3fcb2de0d82b9f6fc0c9a5736
--- .../arch/x86/broadwellx/bdx-metrics.json | 128 ++--- .../pmu-events/arch/x86/broadwellx/cache.json | 88 ++++ .../arch/x86/broadwellx/counter.json | 57 +++ .../arch/x86/broadwellx/floating-point.json | 22 + .../arch/x86/broadwellx/frontend.json | 28 ++ .../arch/x86/broadwellx/memory.json | 58 +++ .../arch/x86/broadwellx/metricgroups.json | 11 + .../pmu-events/arch/x86/broadwellx/other.json | 4 + .../arch/x86/broadwellx/pipeline.json | 137 ++++++ .../arch/x86/broadwellx/uncore-cache.json | 399 +++++++++++++++ .../x86/broadwellx/uncore-interconnect.json | 454 ++++++++++++++++++ .../arch/x86/broadwellx/uncore-io.json | 62 +++ .../arch/x86/broadwellx/uncore-memory.json | 326 +++++++++++++ .../arch/x86/broadwellx/uncore-power.json | 57 +++ .../arch/x86/broadwellx/virtual-memory.json | 38 ++ 15 files changed, 1787 insertions(+), 82 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/broadwellx/counter.json
tools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json +46 -82
@@ -68,7 +68,7 @@ }, { "BriefDescription": "Percentage of time spent in the active CPU power state C0", - "MetricExpr": "tma_info_system_cpu_utilization", + "MetricExpr": "tma_info_system_cpus_utilized", "MetricName": "cpu_utilization", "ScaleUnit": "100%" },
@@ -292,7 +292,7 @@ { "BriefDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists", "MetricExpr": "66 * OTHER_ASSISTS.ANY_WB_ASSIST / tma_info_thread_slots", - "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_group", + "MetricGroup": "BvIO;TopdownL4;tma_L4_group;tma_microcode_sequencer_group", "MetricName": "tma_assists", "MetricThreshold": "tma_assists > 0.1 & (tma_microcode_sequencer > 0.05 & tma_heavy_operations > 0.1)", "PublicDescription": "This metric estimates fraction of slots the CPU retired uops delivered by the Microcode_Sequencer as a result of Assists. Assists are long sequences of uops that are required in certain corner-cases for operations that cannot be handled natively by the execution pipeline. For example; when working with very small floating point values (so-called Denormals); the FP units are not set up to perform these operations natively. Instead; a sequence of instructions to perform the computation on the Denormals is injected into the pipeline. Since these microcode sequences might be dozens of uops long; Assists can be extremely deleterious to performance and they can be avoided in many cases. Sample with: OTHER_ASSISTS.ANY",
@@ -302,7 +302,7 @@ "BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend", "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma_retiring)", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", "MetricgroupNoGroup": "TopdownL1",
@@ -323,7 +323,7 @@ "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Branch Misprediction", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT) * tma_bad_speculation", - "MetricGroup": "BadSpec;BrMispredicts;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -362,7 +362,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(60 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD))) + 43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD)))) / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", "MetricThreshold": "tma_contested_accesses > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to contested accesses. Contested accesses occur when data written by one Logical Processor are read by another Logical Processor on a different Physical Core. Examples of contested accesses include synchronizations such as locks; true data sharing such as modified locked variables; and false sharing. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM_PS;MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS_PS. Related metrics: tma_data_sharing, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -383,7 +383,7 @@ "BriefDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "43 * (MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD))) / tma_info_thread_clks", - "MetricGroup": "Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", + "MetricGroup": "BvMS;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_data_sharing", "MetricThreshold": "tma_data_sharing > 0.05 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles while the memory subsystem was handling synchronizations due to data-sharing accesses. Data shared by multiple Logical Processors (even just read shared) may cause increased access latency due to cache coherency. Excessive data sharing can drastically harm multithreaded performance. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HIT_PS. Related metrics: tma_contested_accesses, tma_false_sharing, tma_machine_clears, tma_remote_cache",
@@ -392,7 +392,7 @@ { "BriefDescription": "This metric represents fraction of cycles where the Divider unit was active", "MetricExpr": "ARITH.FPU_DIV_ACTIVE / tma_info_core_core_clks", - "MetricGroup": "TopdownL3;tma_L3_group;tma_core_bound_group", + "MetricGroup": "BvCB;TopdownL3;tma_L3_group;tma_core_bound_group", "MetricName": "tma_divider", "MetricThreshold": "tma_divider > 0.2 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2)", "PublicDescription": "This metric represents fraction of cycles where the Divider unit was active. Divide and square root instructions are performed by the Divider unit and can take considerably longer latency than integer or Floating Point addition; subtraction; or multiplication. Sample with: ARITH.DIVIDER_UOPS",
@@ -429,7 +429,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses", "MetricExpr": "(8 * DTLB_LOAD_MISSES.STLB_HIT + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_LOAD_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_l1_bound_group", "MetricName": "tma_dtlb_load", "MetricThreshold": "tma_dtlb_load > 0.1 & (tma_l1_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles where the Data TLB (DTLB) was missed by load accesses. TLBs (Translation Look-aside Buffers) are processor caches for recently used entries out of the Page Tables that are used to map virtual- to physical-addresses by the operating system. This metric approximates the potential delay of demand loads missing the first-level data TLB (assuming worst case scenario with back to back misses to different pages). This includes hitting in the second-level TLB (STLB) as well as performing a hardware page walk on an STLB miss. Sample with: MEM_UOPS_RETIRED.STLB_MISS_LOADS_PS. Related metrics: tma_dtlb_store",
@@ -438,7 +438,7 @@ { "BriefDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses", "MetricExpr": "(8 * DTLB_STORE_MISSES.STLB_HIT + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * DTLB_STORE_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", + "MetricGroup": "BvMT;MemoryTLB;TopdownL4;tma_L4_group;tma_issueTLB;tma_store_bound_group", "MetricName": "tma_dtlb_store", "MetricThreshold": "tma_dtlb_store > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates the fraction of cycles spent handling first-level data TLB store misses. As with ordinary data caching; focus on improving data locality and reducing working-set size to reduce DTLB overhead. Additionally; consider using profile-guided optimization (PGO) to collocate frequently-used data on the same page. Try using larger page sizes for large amounts of frequently-used data. Sample with: MEM_UOPS_RETIRED.STLB_MISS_STORES_PS. Related metrics: tma_dtlb_load",
@@ -447,7 +447,7 @@ { "BriefDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing", "MetricExpr": "(200 * OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.REMOTE_HITM + 60 * OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.HITM_OTHER_CORE) / tma_info_thread_clks", - "MetricGroup": "DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", + "MetricGroup": "BvMS;DataSharing;Offcore;Snoop;TopdownL4;tma_L4_group;tma_issueSyncxn;tma_store_bound_group", "MetricName": "tma_false_sharing", "MetricThreshold": "tma_false_sharing > 0.05 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric roughly estimates how often CPU was handling synchronizations due to False Sharing. False Sharing is a multithreading hiccup; where multiple Logical Processors contend on different data-elements mapped into the same cache line. Sample with: MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM_PS;OFFCORE_RESPONSE.DEMAND_RFO.L3_HIT.SNOOP_HITM. Related metrics: tma_contested_accesses, tma_data_sharing, tma_machine_clears, tma_remote_cache",
@@ -457,7 +457,7 @@ "BriefDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_info_memory_load_miss_real_latency * cpu@L1D_PEND_MISS.FB_FULL\\,cmask\\=1@ / tma_info_thread_clks", - "MetricGroup": "MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", + "MetricGroup": "BvMS;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;tma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", "PublicDescription": "This metric does a *rough estimation* of how often L1D Fill Buffer unavailability limited additional L1D miss memory access requests to proceed. The higher the metric value; the deeper the memory hierarchy level the misses are satisfied from (metric values >1 are valid). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or external memory). Related metrics: tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_latency, tma_streaming_stores",
@@ -494,7 +494,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.SCALAR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -503,7 +503,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-point (FP) vector uops fraction the CPU has retired aggregated across all vector widths", - "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@ / UOPS_RETIRED.RETIRE_SLOTS", + "MetricExpr": "FP_ARITH_INST_RETIRED.VECTOR / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6)",
@@ -531,7 +531,7 @@ { "BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend", "MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / tma_info_thread_slots", - "MetricGroup": "PGO;TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvFB;BvIO;PGO;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", "MetricgroupNoGroup": "TopdownL1",
@@ -551,7 +551,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to instruction cache misses.", "MetricExpr": "ICACHE.IFDATA_STALL / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;IcMiss;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_icache_misses", "MetricThreshold": "tma_icache_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "ScaleUnit": "100%"
@@ -590,7 +590,7 @@ }, { "BriefDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width)", - "MetricExpr": "(cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@) / (2 * tma_info_core_core_clks)", + "MetricExpr": "(FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR) / (2 * tma_info_core_core_clks)", "MetricGroup": "Cor;Flops;HPC", "MetricName": "tma_info_core_fp_arith_utilization", "PublicDescription": "Actual per-core usage of the Floating Point non-X87 execution units (regardless of precision or vector-width). Values > 1 are possible due to ([BDW+] Fused-Multiply Add (FMA) counting - common; [ADL+] use all of ADD/MUL/FMA in Scalar or 128/256-bit vectors - less common)."
@@ -630,7 +630,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (lower number means higher occurrence rate)", - "MetricExpr": "INST_RETIRED.ANY / (cpu@FP_ARITH_INST_RETIRED.SCALAR_SINGLE\\,umask\\=0x03@ + cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=0x3c@)", + "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + FP_ARITH_INST_RETIRED.VECTOR)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10",
@@ -704,12 +704,12 @@ "MetricThreshold": "tma_info_inst_mix_ipstore < 8" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Instructions per taken branch", "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW;Frontend;PGO;tma_issueFB", "MetricName": "tma_info_inst_mix_iptb", "MetricThreshold": "tma_info_inst_mix_iptb < 9", - "PublicDescription": "Instruction per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" + "PublicDescription": "Instructions per taken branch. Related metrics: tma_dsb_switches, tma_fetch_bandwidth, tma_info_frontend_dsb_coverage, tma_lcp" }, { "BriefDescription": "Average per-core data fill bandwidth to the L1 data cache [GB / sec]",
@@ -730,23 +730,11 @@ "MetricName": "tma_info_memory_core_l3_cache_fill_bw_2t" }, { - "BriefDescription": "Average Parallel L2 cache miss data reads", - "MetricExpr": "tma_info_memory_latency_data_l2_mlp", - "MetricGroup": "Memory_BW;Offcore", - "MetricName": "tma_info_memory_data_l2_mlp" - }, - { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L1 data cache [GB / sec]", "MetricExpr": "64 * L1D.REPLACEMENT / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw" }, - { - "BriefDescription": "Average per-core data fill bandwidth to the L1 data cache [GB / sec]", - "MetricExpr": "64 * L1D.REPLACEMENT / 1e9 / (duration_time * 1e3 / 1e3)", - "MetricGroup": "Mem;MemoryBW", - "MetricName": "tma_info_memory_l1d_cache_fill_bw_2t" - }, { "BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads", "MetricExpr": "1e3 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
@@ -754,17 +742,11 @@ "MetricName": "tma_info_memory_l1mpki" }, { - "BriefDescription": "", + "BriefDescription": "Average per-thread data fill bandwidth to the L2 cache [GB / sec]", "MetricExpr": "64 * L2_LINES_IN.ALL / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l2_cache_fill_bw" }, - { - "BriefDescription": "Average per-core data fill bandwidth to the L2 cache [GB / sec]", - "MetricExpr": "64 * L2_LINES_IN.ALL / 1e9 / (duration_time * 1e3 / 1e3)", - "MetricGroup": "Mem;MemoryBW", - "MetricName": "tma_info_memory_l2_cache_fill_bw_2t" - }, { "BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)", "MetricExpr": "1e3 * (L2_RQSTS.REFERENCES - L2_RQSTS.MISS) / INST_RETIRED.ANY",
@@ -796,16 +778,16 @@ "MetricName": "tma_info_memory_l2mpki_load" }, { - "BriefDescription": "", - "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1e9 / duration_time", - "MetricGroup": "Mem;MemoryBW", - "MetricName": "tma_info_memory_l3_cache_fill_bw" + "BriefDescription": "Offcore requests (L2 cache miss) per kilo instruction for demand RFOs", + "MetricExpr": "1e3 * OFFCORE_REQUESTS.DEMAND_RFO / INST_RETIRED.ANY", + "MetricGroup": "CacheMisses;Offcore", + "MetricName": "tma_info_memory_l2mpki_rfo" }, { - "BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]", - "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1e9 / (duration_time * 1e3 / 1e3)", + "BriefDescription": "Average per-thread data fill bandwidth to the L3 cache [GB / sec]", + "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1e9 / duration_time", "MetricGroup": "Mem;MemoryBW", - "MetricName": "tma_info_memory_l3_cache_fill_bw_2t" + "MetricName": "tma_info_memory_l3_cache_fill_bw" }, { "BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
@@ -819,29 +801,17 @@ "MetricGroup": "Memory_BW;Offcore", "MetricName": "tma_info_memory_latency_data_l2_mlp" }, - { - "BriefDescription": "Average Latency for L2 cache miss demand Loads", - "MetricExpr": "tma_info_memory_load_l2_miss_latency", - "MetricGroup": "Memory_Lat;Offcore", - "MetricName": "tma_info_memory_latency_load_l2_miss_latency" - }, - { - "BriefDescription": "Average Parallel L2 cache miss demand Loads", - "MetricExpr": "tma_info_memory_load_l2_mlp", - "MetricGroup": "Memory_BW;Offcore", - "MetricName": "tma_info_memory_latency_load_l2_mlp" - }, { "BriefDescription": "Average Latency for L2 cache miss demand Loads", "MetricExpr": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD / OFFCORE_REQUESTS.DEMAND_DATA_RD", "MetricGroup": "Memory_Lat;Offcore", - "MetricName": "tma_info_memory_load_l2_miss_latency" + "MetricName": "tma_info_memory_latency_load_l2_miss_latency" }, { "BriefDescription": "Average Parallel L2 cache miss demand Loads", "MetricExpr": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD / OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD", "MetricGroup": "Memory_BW;Offcore", - "MetricName": "tma_info_memory_load_l2_mlp" + "MetricName": "tma_info_memory_latency_load_l2_mlp" }, { "BriefDescription": "Actual Average Latency for L1 data-cache miss demand load operations (in core cycles)",
@@ -858,12 +828,6 @@ "MetricName": "tma_info_memory_mlp", "PublicDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-Logical Processor)" }, - { - "BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses", - "MetricExpr": "tma_info_memory_tlb_page_walks_utilization", - "MetricGroup": "Mem;MemoryTLB", - "MetricName": "tma_info_memory_page_walks_utilization" - }, { "BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses", "MetricExpr": "(ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION + 7 * (DTLB_STORE_MISSES.WALK_COMPLETED + DTLB_LOAD_MISSES.WALK_COMPLETED + ITLB_MISSES.WALK_COMPLETED)) / (2 * tma_info_core_core_clks)",
@@ -872,7 +836,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0.5" }, { - "BriefDescription": "", + "BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is execution) per core", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute"
@@ -891,13 +855,13 @@ }, { "BriefDescription": "Average CPU Utilization (percentage)", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "tma_info_system_cpus_utilized / #num_cpus_online", "MetricGroup": "HPC;Summary", "MetricName": "tma_info_system_cpu_utilization" }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "#num_cpus_online * tma_info_system_cpu_utilization", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" },
@@ -1012,7 +976,7 @@ "MetricThreshold": "tma_info_thread_uoppi > 1.05" }, { - "BriefDescription": "Instruction per taken branch", + "BriefDescription": "Uops per taken branch", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / BR_INST_RETIRED.NEAR_TAKEN", "MetricGroup": "Branches;Fed;FetchBW", "MetricName": "tma_info_thread_uptb",
@@ -1021,7 +985,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses", "MetricExpr": "(14 * ITLB_MISSES.STLB_HIT + cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * ITLB_MISSES.WALK_COMPLETED) / tma_info_thread_clks", - "MetricGroup": "BigFootprint;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;MemoryTLB;TopdownL3;tma_L3_group;tma_fetch_latency_group", "MetricName": "tma_itlb_misses", "MetricThreshold": "tma_itlb_misses > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15)", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Instruction TLB (ITLB) misses. Sample with: ITLB_MISSES.WALK_COMPLETED",
@@ -1039,7 +1003,7 @@ { "BriefDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks", - "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", + "MetricGroup": "BvML;CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_memory_bound_group", "MetricName": "tma_l2_bound", "MetricThreshold": "tma_l2_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2)", "PublicDescription": "This metric estimates how often the CPU was stalled due to L2 cache accesses by loads. Avoiding cache misses (i.e. L1 misses/L2 hits) can improve the latency and increase performance. Sample with: MEM_LOAD_UOPS_RETIRED.L2_HIT_PS",
@@ -1059,7 +1023,7 @@ "BriefDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited)", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "41 * (MEM_LOAD_UOPS_RETIRED.L3_HIT * (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOAD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD))) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", + "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles with demand load accesses that hit the L3 cache under unloaded scenarios (possibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3 hits) will improve the latency; reduce contention with sibling physical cores and increase performance. Note the value of this node may overlap with its siblings. Sample with: MEM_LOAD_UOPS_RETIRED.L3_HIT_PS. Related metrics: tma_mem_latency",
@@ -1117,7 +1081,7 @@ "BriefDescription": "This metric represents fraction of slots the CPU has wasted due to Machine Clears", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bad_speculation - tma_branch_mispredicts", - "MetricGroup": "BadSpec;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", + "MetricGroup": "BadSpec;BvMS;MachineClears;TmaL2;TopdownL2;tma_L2_group;tma_bad_speculation_group;tma_issueMC;tma_issueSyncxn", "MetricName": "tma_machine_clears", "MetricThreshold": "tma_machine_clears > 0.1 & tma_bad_speculation > 0.15", "MetricgroupNoGroup": "TopdownL2",
@@ -1127,7 +1091,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, cpu@OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD\\,cmask\\=4@) / tma_info_thread_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the core's performance was likely hurt due to approaching bandwidth limits of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuristic assumes that a similar off-core traffic is generated by all IA cores. This metric does not aggregate non-data-read requests by this logical processor; requests from other IA Logical Processors/Physical Cores/sockets; or other non-IA devices like GPU; hence the maximum external memory bandwidth limits may or may not be approached when this metric is flagged (see Uncore counters for that). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full",
@@ -1136,7 +1100,7 @@ { "BriefDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM)", "MetricExpr": "min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD) / tma_info_thread_clks - tma_mem_bandwidth", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles where the performance was likely hurt due to latency from external memory - DRAM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from other Logical Processors/Physical Cores/sockets (see Uncore counters for that). Related metrics: tma_l3_hit_latency",
@@ -1165,7 +1129,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES * tma_branch_resteers / (BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY)", - "MetricGroup": "BadSpec;BrMispredicts;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", + "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", "MetricThreshold": "tma_mispredicts_resteers > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to Branch Resteers as a result of Branch Misprediction at execution stage. Related metrics: tma_branch_mispredicts, tma_info_bad_spec_branch_misprediction_cost",
@@ -1301,7 +1265,7 @@ { "BriefDescription": "This metric represents fraction of cycles CPU executed total of 3 or more uops per cycle on all execution ports (Logical Processor cycles since ICL, Physical Core cycles otherwise).", "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,cmask\\=3@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks", - "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", + "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_utilization_group", "MetricName": "tma_ports_utilized_3m", "MetricThreshold": "tma_ports_utilized_3m > 0.4 & (tma_ports_utilization > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", "ScaleUnit": "100%"
@@ -1328,7 +1292,7 @@ { "BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / tma_info_thread_slots", - "MetricGroup": "TmaL1;TopdownL1;tma_L1_group", + "MetricGroup": "BvUW;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.1", "MetricgroupNoGroup": "TopdownL1",
@@ -1357,7 +1321,7 @@ { "BriefDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors)", "MetricExpr": "(OFFCORE_REQUESTS_BUFFER.SQ_FULL / 2 if #SMT_on else OFFCORE_REQUESTS_BUFFER.SQ_FULL) / tma_info_core_core_clks", - "MetricGroup": "MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", + "MetricGroup": "BvMS;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_issueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric measures fraction of cycles where the Super Queue (SQ) was full taking into account all request-types and both hardware SMT threads (Logical Processors). Related metrics: tma_fb_full, tma_info_system_dram_bw_use, tma_mem_bandwidth",
@@ -1385,7 +1349,7 @@ "BriefDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses", "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(L2_RQSTS.RFO_HIT * 9 * (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) + (1 - MEM_UOPS_RETIRED.LOCK_LOADS / MEM_UOPS_RETIRED.ALL_STORES) * min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO)) / tma_info_thread_clks", - "MetricGroup": "MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", + "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_issueRFO;tma_issueSL;tma_store_bound_group", "MetricName": "tma_store_latency", "MetricThreshold": "tma_store_latency > 0.1 & (tma_store_bound > 0.2 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", "PublicDescription": "This metric estimates fraction of cycles the CPU spent handling L1D store misses. Store accesses usually less impact out-of-order core performance; however; holding resources for longer time can lead into undesired implications (e.g. contention on L1D fill-buffer entries - see FB_Full). Related metrics: tma_fb_full, tma_lock_latency",
@@ -1402,7 +1366,7 @@ { "BriefDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears", "MetricExpr": "tma_branch_resteers - tma_mispredicts_resteers - tma_clears_resteers", - "MetricGroup": "BigFootprint;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", + "MetricGroup": "BigFootprint;BvBC;FetchLat;TopdownL4;tma_L4_group;tma_branch_resteers_group", "MetricName": "tma_unknown_branches", "MetricThreshold": "tma_unknown_branches > 0.05 & (tma_branch_resteers > 0.05 & (tma_fetch_latency > 0.1 & tma_frontend_bound > 0.15))", "PublicDescription": "This metric represents fraction of cycles the CPU was stalled due to new branch address clears. These are fetched branches the Branch Prediction Unit was unable to recognize (e.g. first time the branch is fetched or hitting BTB capacity limit) hence called Unknown Branches. Sample with: BACLEARS.ANY",
tools/perf/pmu-events/arch/x86/broadwellx/cache.json +88 -0
616 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/counter.json new file +57 -0
58 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/floating-point.json +22 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -17,6 +19,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 4 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 4 computation operations, one for each element. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -25,6 +28,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 8 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", "PublicDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 8 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -33,6 +37,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational 128-bit packed single and 256-bit packed double precision FP instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, 1 for each element. Applies to SSE* and AVX* packed single precision and packed double precision FP instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB count twice as they perform 2 calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", "PublicDescription": "Number of SSE/AVX computational 128-bit packed single precision and 256-bit packed double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 2 or/and 4 computation operations, one for each element. Applies to SSE* and AVX* packed single precision floating-point and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -41,6 +46,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.DOUBLE", "SampleAfterValue": "2000006",
@@ -48,6 +54,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* packed double and single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.PACKED", "SampleAfterValue": "2000004",
@@ -55,6 +62,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computation operation. Applies to SSE* and AVX* scalar double and single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", "PublicDescription": "Number of SSE/AVX computational scalar single precision and double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -63,6 +71,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", "PublicDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -71,6 +80,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", "PublicDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired; some instructions will count twice as noted below. Each count represents 1 computational operation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR register need to be set when using these events.",
@@ -79,6 +89,7 @@ }, { "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired; some instructions will count twice as noted below. Applies to SSE* and AVX* scalar and packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SINGLE", "SampleAfterValue": "2000005",
@@ -86,6 +97,7 @@ }, { "BriefDescription": "Number of any Vector retired FP arithmetic instructions", + "Counter": "0,1,2,3", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", "SampleAfterValue": "2000003",
@@ -93,6 +105,7 @@ }, { "BriefDescription": "Cycles with any input/output SSE or FP assist", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0xCA", "EventName": "FP_ASSIST.ANY",
@@ -102,6 +115,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to input values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_INPUT", "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
@@ -110,6 +124,7 @@ }, { "BriefDescription": "Number of SIMD FP assists due to Output values", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.SIMD_OUTPUT", "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
@@ -118,6 +133,7 @@ }, { "BriefDescription": "Number of X87 assists due to input value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_INPUT", "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
@@ -126,6 +142,7 @@ }, { "BriefDescription": "Number of X87 assists due to output value.", + "Counter": "0,1,2,3", "EventCode": "0xCA", "EventName": "FP_ASSIST.X87_OUTPUT", "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
@@ -134,6 +151,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_ELIMINATED", "SampleAfterValue": "1000003",
@@ -141,6 +159,7 @@ }, { "BriefDescription": "Number of SIMD Move Elimination candidate uops that were not eliminated.", + "Counter": "0,1,2,3", "EventCode": "0x58", "EventName": "MOVE_ELIMINATION.SIMD_NOT_ELIMINATED", "SampleAfterValue": "1000003",
@@ -148,6 +167,7 @@ }, { "BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.AVX_TO_SSE",
@@ -157,6 +177,7 @@ }, { "BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.", + "Counter": "0,1,2,3", "Errata": "BDM30", "EventCode": "0xC1", "EventName": "OTHER_ASSISTS.SSE_TO_AVX",
@@ -166,6 +187,7 @@ }, { "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports", + "Counter": "0,1,2,3", "EventCode": "0xA0", "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF", "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
tools/perf/pmu-events/arch/x86/broadwellx/frontend.json +28 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.", + "Counter": "0,1,2,3", "EventCode": "0xe6", "EventName": "BACLEARS.ANY", "SampleAfterValue": "100003",
@@ -8,6 +9,7 @@ }, { "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.", + "Counter": "0,1,2,3", "EventCode": "0xAB", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
@@ -16,6 +18,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Reads. both cacheable and noncacheable, including UC fetches", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.HIT", "PublicDescription": "This event counts the number of both cacheable and noncacheable Instruction Cache, Streaming Buffer and Victim Cache Reads including UC fetches.",
@@ -24,6 +27,7 @@ }, { "BriefDescription": "Cycles where a code fetch is stalled due to L1 instruction-cache miss.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.IFDATA_STALL", "PublicDescription": "This event counts cycles during which the demand fetch waits for data (wfdM104H) from L2 or iSB (opportunistic hit).",
@@ -32,6 +36,7 @@ }, { "BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Misses. Includes Uncacheable accesses.", + "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE.MISSES", "PublicDescription": "This event counts the number of instruction cache, streaming buffer and victim cache misses. Counting includes UC accesses.",
@@ -40,6 +45,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
@@ -49,6 +55,7 @@ }, { "BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
@@ -58,6 +65,7 @@ }, { "BriefDescription": "Cycles MITE is delivering 4 Uops", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
@@ -67,6 +75,7 @@ }, { "BriefDescription": "Cycles MITE is delivering any Uop", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
@@ -76,6 +85,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES",
@@ -85,6 +95,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
@@ -93,6 +104,7 @@ }, { "BriefDescription": "Instruction Decode Queue (IDQ) empty cycles", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.EMPTY", "PublicDescription": "This counts the number of cycles that the instruction decoder queue is empty and can indicate that the application may be bound in the front end. It does not determine whether there are uops being delivered to the Alloc stage since uops can be delivered by bypass skipping the Instruction Decode Queue (IDQ) when it is empty.",
@@ -101,6 +113,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_ALL_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -109,6 +122,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES",
@@ -118,6 +132,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
@@ -126,6 +141,7 @@ }, { "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES",
@@ -135,6 +151,7 @@ }, { "BriefDescription": "Cycles when uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_CYCLES",
@@ -144,6 +161,7 @@ }, { "BriefDescription": "Deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -154,6 +172,7 @@ }, { "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_DSB_UOPS", "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -162,6 +181,7 @@ }, { "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_MITE_UOPS", "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
@@ -170,6 +190,7 @@ }, { "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x79",
@@ -179,6 +200,7 @@ }, { "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequencer (MS) is busy", + "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
@@ -187,6 +209,7 @@ }, { "BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled", + "Counter": "0,1,2,3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
@@ -195,6 +218,7 @@ }, { "BriefDescription": "Cycles per thread when 4 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "4", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE",
@@ -204,6 +228,7 @@ }, { "BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
@@ -213,6 +238,7 @@ }, { "BriefDescription": "Cycles per thread when 3 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled", + "Counter": "0,1,2,3", "CounterMask": "3", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_1_UOP_DELIV.CORE",
@@ -222,6 +248,7 @@ }, { "BriefDescription": "Cycles with less than 2 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "2", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_2_UOP_DELIV.CORE",
@@ -230,6 +257,7 @@ }, { "BriefDescription": "Cycles with less than 3 uops delivered by the front end.", + "Counter": "0,1,2,3", "CounterMask": "1", "EventCode": "0x9C", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
tools/perf/pmu-events/arch/x86/broadwellx/memory.json +58 -0
406 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/metricgroups.json +11 -0
@@ -5,7 +5,18 @@ "BigFootprint": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "BrMispredicts": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Branches": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvBC": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvCB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvFB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvIO": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvML": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMP": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMS": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvMT": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvOB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "BvUW": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "CacheHits": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", + "CacheMisses": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Compute": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "Cor": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet", "DSB": "Grouping from Top-down Microarchitecture Analysis Metrics spreadsheet",
tools/perf/pmu-events/arch/x86/broadwellx/other.json +4 -0
@@ -1,6 +1,7 @@ [ { "BriefDescription": "Unhalted core cycles when the thread is in ring 0", + "Counter": "0,1,2,3", "EventCode": "0x5C", "EventName": "CPL_CYCLES.RING0", "PublicDescription": "This event counts the unhalted core cycles during which the thread is in the ring 0 privileged mode.",
@@ -9,6 +10,7 @@ }, { "BriefDescription": "Number of intervals between processor halts while thread is in ring 0", + "Counter": "0,1,2,3", "CounterMask": "1", "EdgeDetect": "1", "EventCode": "0x5C",
@@ -19,6 +21,7 @@ }, { "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3", + "Counter": "0,1,2,3", "EventCode": "0x5C", "EventName": "CPL_CYCLES.RING123", "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
@@ -27,6 +30,7 @@ }, { "BriefDescription": "Cycles when L1 and L2 are locked due to UC or split lock", + "Counter": "0,1,2,3", "EventCode": "0x63", "EventName": "LOCK_CYCLES.SPLIT_LOCK_UC_LOCK_DURATION", "PublicDescription": "This event counts cycles in which the L1 and L2 are locked due to a UC lock or split lock. A lock is asserted in case of locked memory access, due to noncacheable memory, locked operation that spans two cache lines, or a page walk from the noncacheable page table. L1D and L2 locks have a very high performance penalty and it is highly recommended to avoid such access.",
tools/perf/pmu-events/arch/x86/broadwellx/pipeline.json +137 -0
959 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/uncore-cache.json +399 -0
180 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/uncore-interconnect.json +454 -0
454 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/uncore-io.json +62 -0
62 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/uncore-memory.json +326 -0
326 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json +57 -0
57 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/broadwellx/virtual-memory.json +38 -0
Lines of this file are not shown. View the raw message.
The rest of this diff is not shown. View the raw message.
Signature · 1 line
2.45.2.627.g7a2c4fd464-goog
Part of the diff is not shown here. View the raw message.