Skip to contentSkip to conversation

perf vendor events intel: Add core events and metrics for Elkhartlake

PATCH Applied Jin Yao 2 participants 4 messages Aug 2, 2021 perf events

lore
Share & tools
lore

Jin Yao

author
#0 lore raw

v2 cover perf vendor events intel: Add core events and metrics for Elkhartlake

Add JSON events and metrics for Elkhartlake to perf.

Based on JSON list v1.02:

https://download.01.org/perfmon/EHL/

v2:
---
Based on JSON list v1.02.
Patchset is based on tmp.perf/core branch.

v1:
---
Based on JSON list v1.01

Jin Yao (2):
perf vendor events intel: Add core event list for Elkhartlake perf vendor events intel: Add basic metrics for Elkhartlake

 .../arch/x86/elkhartlake/cache.json           | 226 ++++++++++
 .../arch/x86/elkhartlake/ehl-metrics.json     |  57 +++
 .../arch/x86/elkhartlake/floating-point.json  |  24 +
 .../arch/x86/elkhartlake/frontend.json        |  93 ++++
 .../arch/x86/elkhartlake/memory.json          |  86 ++++
 .../arch/x86/elkhartlake/other.json           | 424 ++++++++++++++++++
 .../arch/x86/elkhartlake/pipeline.json        | 278 ++++++++++++
 .../arch/x86/elkhartlake/virtual-memory.json  | 273 +++++++++++
 tools/perf/pmu-events/arch/x86/mapfile.csv    |   1 +
 9 files changed, 1462 insertions(+)
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/cache.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/ehl-metrics.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/floating-point.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/frontend.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/memory.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/other.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/pipeline.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/virtual-memory.json
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH v2 1/2 perf vendor events intel: Add core event list for Elkhartlake

Add JSON core events for Elkhartlake to perf.

Based on JSON list v1.02:

https://download.01.org/perfmon/EHL/

---
v2:
- Based on JSON list v1.02.

v1:
- Based on JSON list v1.01.

 .../arch/x86/elkhartlake/cache.json           | 226 ++++++++++
 .../arch/x86/elkhartlake/floating-point.json  |  24 +
 .../arch/x86/elkhartlake/frontend.json        |  93 ++++
 .../arch/x86/elkhartlake/memory.json          |  86 ++++
 .../arch/x86/elkhartlake/other.json           | 424 ++++++++++++++++++
 .../arch/x86/elkhartlake/pipeline.json        | 278 ++++++++++++
 .../arch/x86/elkhartlake/virtual-memory.json  | 273 +++++++++++
 tools/perf/pmu-events/arch/x86/mapfile.csv    |   1 +
 8 files changed, 1405 insertions(+)
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/cache.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/floating-point.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/frontend.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/memory.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/other.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/pipeline.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/virtual-memory.json
tools/perf/pmu-events/arch/x86/elkhartlake/cache.json new file +226 -0
new file mode 100644
index 000000000000..734be4ea095f
@@ -0,0 +1,226 @@
+[
+    {
+        "BriefDescription": "Counts the number of first level data cacheline (dirty) evictions caused by misses, stores, and prefetches.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x51",
+        "EventName": "DL1.DIRTY_EVICTION",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of first level data cacheline (dirty) evictions caused by misses, stores, and prefetches.  Does not count evictions or dirty writebacks caused by snoops.  Does not count a replacement unless a (dirty) line was written back.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x1"
+    },
+    {
+        "BriefDescription": "Counts the number of cacheable memory requests that miss in the LLC. Counts on a per core basis.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x2e",
+        "EventName": "LONGEST_LAT_CACHE.MISS",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cacheable memory requests that miss in the Last Level Cache (LLC). If the platform has an L3 cache, the LLC is the L3 cache, otherwise it is the L2 cache. Counts on a per core basis.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x41"
+    },
+    {
+        "BriefDescription": "Counts the number of cacheable memory requests that access the LLC. Counts on a per core basis.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x2e",
+        "EventName": "LONGEST_LAT_CACHE.REFERENCE",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cacheable memory requests that access the Last Level Cache (LLC). Requests include demand loads, reads for ownership (RFO), instruction fetches and L1 HW prefetches. If the platform has an L3 cache, the LLC is the L3 cache, otherwise it is the L2 cache. Counts on a per core basis.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x4f"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to an instruction cache or TLB miss which hit in DRAM or MMIO (Non-DRAM).",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.IFETCH_DRAM_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cycles a core is stalled due to an instruction cache or translation lookaside buffer (TLB) access which hit in DRAM or MMIO (non-DRAM).",
+        "SampleAfterValue": "200003",
+        "UMask": "0x20"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to an instruction cache or TLB miss which hit in the L2 cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.IFETCH_L2_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cycles a core is stalled due to an instruction cache or Translation Lookaside Buffer (TLB) access which hit in the L2 cache.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x8"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to an instruction cache or TLB miss which hit in the LLC or other core with HITE/F/M.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.IFETCH_LLC_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cycles a core is stalled due to an instruction cache or Translation Lookaside Buffer (TLB) access which hit in the Last Level Cache (LLC) or other core with HITE/F/M.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x10"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load miss which hit in DRAM or MMIO (Non-DRAM).",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.LOAD_DRAM_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x4"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load which hit in the L2 cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.LOAD_L2_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the number of cycles a core is stalled due to a demand load which hit in the L2 cache.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x1"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles the core is stalled due to a demand load which hit in the LLC or other core with HITE/F/M.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.LOAD_LLC_HIT",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "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.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x2"
+    },
+    {
+        "BriefDescription": "Counts the number of cycles a core is stalled due to a store buffer being full.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x34",
+        "EventName": "MEM_BOUND_STALLS.STORE_BUFFER_FULL",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x40"
+    },
+    {
+        "BriefDescription": "Counts the number of load ops retired that hit in DRAM.",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.DRAM_HIT",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x80"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired that hit in the L1 data cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x1"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired that miss in the L1 data cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x8"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired that hit in the L2 cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x2"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired that miss in the L2 cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x10"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired that hit in the L3 cache.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0xd1",
+        "EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x4"
+    },
+    {
+        "BriefDescription": "Counts the number of load uops retired.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd0",
+        "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the total number of load uops retired.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x81"
+    },
+    {
+        "BriefDescription": "Counts the number of store uops retired.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "Data_LA": "1",
+        "EventCode": "0xd0",
+        "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "PublicDescription": "Counts the total number of store uops retired.",
+        "SampleAfterValue": "200003",
+        "UMask": "0x82"
+    },
+    {
+        "BriefDescription": "Counts the number of issue slots every cycle that were not delivered by the frontend due to instruction cache misses.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0x71",
+        "EventName": "TOPDOWN_FE_BOUND.ICACHE",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "1000003",
+        "UMask": "0x20"
+    }
+]
\ No newline at end of file
tools/perf/pmu-events/arch/x86/elkhartlake/floating-point.json new file +24 -0
new file mode 100644
index 000000000000..2515b9aa6e66
@@ -0,0 +1,24 @@
+[
+    {
+        "BriefDescription": "Counts the number of cycles the floating point divider is busy.  Does not imply a stall waiting for the divider.",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0xcd",
+        "EventName": "CYCLES_DIV_BUSY.FPDIV",
+        "PDIR_COUNTER": "na",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "200003",
+        "UMask": "0x2"
+    },
+    {
+        "BriefDescription": "Counts the number of floating point divide uops retired (x87 and SSE, including x87 sqrt).",
+        "CollectPEBSRecord": "2",
+        "Counter": "0,1,2,3",
+        "EventCode": "0xc2",
+        "EventName": "UOPS_RETIRED.FPDIV",
+        "PEBS": "1",
+        "PEBScounters": "0,1,2,3",
+        "SampleAfterValue": "2000003",
+        "UMask": "0x8"
+    }
+]
\ No newline at end of file
tools/perf/pmu-events/arch/x86/elkhartlake/frontend.json new file +93 -0

94 lines not on this page · raw message

tools/perf/pmu-events/arch/x86/elkhartlake/memory.json new file +86 -0

87 lines not on this page · raw message

tools/perf/pmu-events/arch/x86/elkhartlake/other.json new file +424 -0

425 lines not on this page · raw message

tools/perf/pmu-events/arch/x86/elkhartlake/pipeline.json new file +278 -0

279 lines not on this page · raw message

tools/perf/pmu-events/arch/x86/elkhartlake/virtual-memory.json new file +273 -0

274 lines not on this page · raw message

tools/perf/pmu-events/arch/x86/mapfile.csv +1 -0
index 37382343d25f..0cf2d1fa6b76 100644
@@ -41,6 +41,7 @@ GenuineIntel-6-A7,v1,icelake,core
 GenuineIntel-6-6A,v1,icelakex,core
 GenuineIntel-6-6C,v1,icelakex,core
 GenuineIntel-6-86,v1,tremontx,core
+GenuineIntel-6-96,v1,elkhartlake,core
 AuthenticAMD-23-([12][0-9A-F]|[0-9A-F]),v2,amdzen1,core
 AuthenticAMD-23-[[:xdigit:]]+,v1,amdzen2,core
 AuthenticAMD-25-[[:xdigit:]]+,v1,amdzen3,core
Signature · 1 line
2.17.1
author
#2 lore raw

PATCH v2 2/2 perf vendor events intel: Add basic metrics for Elkhartlake

In reply to Jin Yao

Add JSON metrics for Elkhartlake to perf.

---
v2:
- No change.

 .../arch/x86/elkhartlake/ehl-metrics.json     | 57 +++++++++++++++++++
 1 file changed, 57 insertions(+)
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/ehl-metrics.json
tools/perf/pmu-events/arch/x86/elkhartlake/ehl-metrics.json new file +57 -0
new file mode 100644
index 000000000000..b6f7126be1fd
@@ -0,0 +1,57 @@
+[
+    {
+        "MetricExpr": "INST_RETIRED.ANY / cycles",
+        "BriefDescription": "Instructions Per Cycle (per Logical Processor)",
+        "MetricName": "IPC"
+    },
+    {
+        "MetricExpr": "1 / IPC",
+        "BriefDescription": "Cycles Per Instruction (per Logical Processor)",
+        "MetricName": "CPI"
+    },
+    {
+        "MetricExpr": "cycles",
+        "BriefDescription": "Per-Logical Processor actual clocks when the Logical Processor is active.",
+        "MetricName": "CLKS"
+    },
+    {
+        "MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
+        "BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
+        "MetricName": "IpMispredict"
+    },
+    {
+        "MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
+        "BriefDescription": "Instructions per Branch (lower number means higher occurrence rate)",
+        "MetricName": "IpBranch"
+    },
+    {
+        "MetricExpr": "INST_RETIRED.ANY",
+        "BriefDescription": "Total number of retired Instructions",
+        "MetricName": "Instructions"
+    },
+    {
+        "MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 ",
+        "BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
+        "MetricName": "L3_Cache_Fill_BW"
+    },
+    {
+        "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
+        "BriefDescription": "Average CPU Utilization",
+        "MetricName": "CPU_Utilization"
+    },
+    {
+        "MetricExpr": "(cycles / CPU_CLK_UNHALTED.REF_TSC) * msr@tsc@ / 1000000000 ",
+        "BriefDescription": "Measured Average Frequency for unhalted processors [GHz]",
+        "MetricName": "Average_Frequency"
+    },
+    {
+        "MetricExpr": "cycles / CPU_CLK_UNHALTED.REF_TSC",
+        "BriefDescription": "Average Frequency Utilization relative nominal frequency",
+        "MetricName": "Turbo_Utilization"
+    },
+    {
+        "MetricExpr": "cycles:k / cycles",
+        "BriefDescription": "Fraction of cycles spent in the Operating System (OS) Kernel mode",
+        "MetricName": "Kernel_Utilization"
+    }
+]
Signature · 1 line
2.17.1
maintainer Applied
#3 lore raw
In reply to Jin Yao

Em Mon, Aug 02, 2021 at 01:34:38PM +0800, Jin Yao escreveu:

Quoted 5 lines

Add JSON events and metrics for Elkhartlake to perf.

Based on JSON list v1.02:

https://download.01.org/perfmon/EHL/

Thanks, applied.

- Arnaldo

Quoted 35 lines

v2:
---
Based on JSON list v1.02.
Patchset is based on tmp.perf/core branch.

v1:
---
Based on JSON list v1.01

Jin Yao (2):
perf vendor events intel: Add core event list for Elkhartlake perf vendor events intel: Add basic metrics for Elkhartlake

 .../arch/x86/elkhartlake/cache.json           | 226 ++++++++++
 .../arch/x86/elkhartlake/ehl-metrics.json     |  57 +++
 .../arch/x86/elkhartlake/floating-point.json  |  24 +
 .../arch/x86/elkhartlake/frontend.json        |  93 ++++
 .../arch/x86/elkhartlake/memory.json          |  86 ++++
 .../arch/x86/elkhartlake/other.json           | 424 ++++++++++++++++++
 .../arch/x86/elkhartlake/pipeline.json        | 278 ++++++++++++
 .../arch/x86/elkhartlake/virtual-memory.json  | 273 +++++++++++
 tools/perf/pmu-events/arch/x86/mapfile.csv    |   1 +
 9 files changed, 1462 insertions(+)
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/cache.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/ehl-metrics.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/floating-point.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/frontend.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/memory.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/other.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/pipeline.json
 create mode 100644 tools/perf/pmu-events/arch/x86/elkhartlake/virtual-memory.json
Signature · 1 line
2.17.1
Signature · 1 line
- Arnaldo