Python generated AMD Zen metrics
Ian Rogers
reviewerv2 cover Python generated AMD Zen metrics
Generate thirteen sets of additional metrics for AMD zen. Rapl and Idle metrics aren't specific to AMD but are placed here for ease and convenience. Uncore L3 metrics are added along with the majority of core metrics.
The patches should be applied on top of:
https://lore.kernel.org/lkml/[email protected]/
v2. Drop the cycles breakdown in favor of having it as a common
metric, suggested by Kan Liang <[email protected]>.
Ian Rogers (12):
perf jevents: Add RAPL event metric for AMD zen models perf jevents: Add idle metric for AMD zen models
perf jevents: Add upc metric for uops per cycle for AMD perf jevents: Add br metric group for branch statistics on AMD perf jevents: Add software prefetch (swpf) metric group for AMD perf jevents: Add hardware prefetch (hwpf) metric group for AMD perf jevents: Add itlb metric group for AMD
perf jevents: Add dtlb metric group for AMD
perf jevents: Add uncore l3 metric group for AMD
perf jevents: Add load store breakdown metrics ldst for AMD perf jevents: Add ILP metrics for AMD
perf jevents: Add context switch metrics for AMD
tools/perf/pmu-events/amd_metrics.py | 623 ++++++++++++++++++++++++++- 1 file changed, 620 insertions(+), 3 deletions(-)
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 1/12 perf jevents: Add RAPL event metric for AMD zen models
Add power per second metrics based on RAPL.
--- tools/perf/pmu-events/amd_metrics.py | 30 ++++++++++++++++++++++++++-- 1 file changed, 28 insertions(+), 2 deletions(-)
tools/perf/pmu-events/amd_metrics.py +28 -2
@@ -1,9 +1,10 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) -from metric import (JsonEncodeMetric, JsonEncodeMetricGroupDescriptions, LoadEvents, - MetricGroup) +from metric import (d_ratio, has_event, Event, JsonEncodeMetric, JsonEncodeMetricGroupDescriptions, + LoadEvents, Metric, MetricGroup, Select) import argparse import json +import math import os parser = argparse.ArgumentParser(description="AMD perf json generator")
@@ -16,6 +17,31 @@ LoadEvents(directory) all_metrics = MetricGroup("",[]) +interval_sec = Event("duration_time") + +def Rapl() -> MetricGroup: + """Processor socket power consumption estimate. + + Use events from the running average power limit (RAPL) driver. + """ + # Watts = joules/second + # Currently only energy-pkg is supported by AMD: + # https://lore.kernel.org/lkml/[email protected]/ + pkg = Event("power/energy\-pkg/") + cond_pkg = Select(pkg, has_event(pkg), math.nan) + scale = 2.3283064365386962890625e-10 + metrics = [ + Metric("cpu_power_pkg", "", + d_ratio(cond_pkg * scale, interval_sec), "Watts"), + ] + + return MetricGroup("cpu_power", metrics, + description="Processor socket power consumption estimates") + +all_metrics = MetricGroup("", [ + Rapl(), +]) + if args.metricgroups: print(JsonEncodeMetricGroupDescriptions(all_metrics)) else:
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 2/12 perf jevents: Add idle metric for AMD zen models
In reply to Ian RogersCompute using the msr PMU the percentage of wallclock cycles where the CPUs are in a low power state.
--- tools/perf/pmu-events/amd_metrics.py | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-)
tools/perf/pmu-events/amd_metrics.py +14 -2
@@ -1,7 +1,8 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) -from metric import (d_ratio, has_event, Event, JsonEncodeMetric, JsonEncodeMetricGroupDescriptions, - LoadEvents, Metric, MetricGroup, Select) +from metric import (d_ratio, has_event, max, Event, JsonEncodeMetric, + JsonEncodeMetricGroupDescriptions, LoadEvents, Metric, + MetricGroup, Select) import argparse import json import math
@@ -19,6 +20,16 @@ all_metrics = MetricGroup("",[]) interval_sec = Event("duration_time") +def Idle() -> Metric: + cyc = Event("msr/mperf/") + tsc = Event("msr/tsc/") + low = max(tsc - cyc, 0) + return Metric( + "idle", + "Percentage of total wallclock cycles where CPUs are in low power state (C1 or deeper sleep state)", + d_ratio(low, tsc), "100%") + + def Rapl() -> MetricGroup: """Processor socket power consumption estimate.
@@ -39,6 +50,7 @@ def Rapl() -> MetricGroup: description="Processor socket power consumption estimates") all_metrics = MetricGroup("", [ + Idle(), Rapl(), ])
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 3/12 perf jevents: Add upc metric for uops per cycle for AMD
In reply to Ian RogersThe metric adjusts for whether or not SMT is on.
--- tools/perf/pmu-events/amd_metrics.py | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-)
tools/perf/pmu-events/amd_metrics.py +13 -4
@@ -1,8 +1,8 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) from metric import (d_ratio, has_event, max, Event, JsonEncodeMetric, - JsonEncodeMetricGroupDescriptions, LoadEvents, Metric, - MetricGroup, Select) + JsonEncodeMetricGroupDescriptions, Literal, LoadEvents, + Metric, MetricGroup, Select) import argparse import json import math
@@ -16,9 +16,17 @@ args = parser.parse_args() directory = f"{os.path.dirname(os.path.realpath(__file__))}/arch/x86/{args.model}/" LoadEvents(directory) -all_metrics = MetricGroup("",[]) - +zen_model = int(args.model[6:]) interval_sec = Event("duration_time") +cycles = Event("cycles") +# Number of CPU cycles scaled for SMT. +smt_cycles = Select(cycles / 2, Literal("#smt_on"), cycles) + +def AmdUpc() -> Metric: + ops = Event("ex_ret_ops", "ex_ret_cops") + upc = d_ratio(ops, smt_cycles) + return Metric("upc", "Micro-ops retired per core cycle (higher is better)", + upc, "uops/cycle") def Idle() -> Metric: cyc = Event("msr/mperf/")
@@ -50,6 +58,7 @@ def Rapl() -> MetricGroup: description="Processor socket power consumption estimates") all_metrics = MetricGroup("", [ + AmdUpc(), Idle(), Rapl(), ])
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 4/12 perf jevents: Add br metric group for branch statistics on AMD
In reply to Ian RogersThe br metric group for branches itself comprises metric groups for total, taken, conditional, fused and far metric groups using json events. The lack of conditional events on anything but zen2 means this category is lacking on zen1, zen3 and zen4.
--- tools/perf/pmu-events/amd_metrics.py | 104 +++++++++++++++++++++++++++ 1 file changed, 104 insertions(+)
tools/perf/pmu-events/amd_metrics.py +104 -0
@@ -7,6 +7,7 @@ import argparse import json import math import os +from typing import Optional parser = argparse.ArgumentParser(description="AMD perf json generator") parser.add_argument("-metricgroups", help="Generate metricgroups data", action='store_true')
@@ -18,10 +19,112 @@ LoadEvents(directory) zen_model = int(args.model[6:]) interval_sec = Event("duration_time") +ins = Event("instructions") cycles = Event("cycles") # Number of CPU cycles scaled for SMT. smt_cycles = Select(cycles / 2, Literal("#smt_on"), cycles) +def AmdBr(): + def Total() -> MetricGroup: + br = Event("ex_ret_brn") + br_m_all = Event("ex_ret_brn_misp") + br_clr = Event("ex_ret_msprd_brnch_instr_dir_msmtch", "ex_ret_brn_resync") + + br_r = d_ratio(br, interval_sec) + ins_r = d_ratio(ins, br) + misp_r = d_ratio(br_m_all, br) + clr_r = d_ratio(br_clr, interval_sec) + + return MetricGroup("br_total", [ + Metric("br_total_retired", + "The number of branch instructions retired per second.", br_r, + "insn/s"), + Metric( + "br_total_mispred", + "The number of branch instructions retired, of any type, that were " + "not correctly predicted as a percentage of all branch instrucions.", + misp_r, "100%"), + Metric("br_total_insn_between_branches", + "The number of instructions divided by the number of branches.", + ins_r, "insn"), + Metric("br_total_insn_fe_resteers", + "The number of resync branches per second.", clr_r, "req/s") + ]) + + def Taken() -> MetricGroup: + br = Event("ex_ret_brn_tkn") + br_m_tk = Event("ex_ret_brn_tkn_misp") + br_r = d_ratio(br, interval_sec) + ins_r = d_ratio(ins, br) + misp_r = d_ratio(br_m_tk, br) + return MetricGroup("br_taken", [ + Metric("br_taken_retired", + "The number of taken branches that were retired per second.", + br_r, "insn/s"), + Metric( + "br_taken_mispred", + "The number of retired taken branch instructions that were " + "mispredicted as a percentage of all taken branches.", misp_r, + "100%"), + Metric( + "br_taken_insn_between_branches", + "The number of instructions divided by the number of taken branches.", + ins_r, "insn"), + ]) + + def Conditional() -> Optional[MetricGroup]: + br = Event("ex_ret_cond") + br_r = d_ratio(br, interval_sec) + ins_r = d_ratio(ins, br) + + metrics = [ + Metric("br_cond_retired", "Retired conditional branch instructions.", + br_r, "insn/s"), + Metric("br_cond_insn_between_branches", + "The number of instructions divided by the number of conditional " + "branches.", ins_r, "insn"), + ] + if zen_model == 2: + br_m_cond = Event("ex_ret_cond_misp") + misp_r = d_ratio(br_m_cond, br) + metrics += [ + Metric("br_cond_mispred", + "Retired conditional branch instructions mispredicted as a " + "percentage of all conditional branches.", misp_r, "100%"), + ] + + return MetricGroup("br_cond", metrics) + + def Fused() -> MetricGroup: + br = Event("ex_ret_fused_instr", "ex_ret_fus_brnch_inst") + br_r = d_ratio(br, interval_sec) + ins_r = d_ratio(ins, br) + return MetricGroup("br_cond", [ + Metric("br_fused_retired", + "Retired fused branch instructions per second.", br_r, "insn/s"), + Metric( + "br_fused_insn_between_branches", + "The number of instructions divided by the number of fused " + "branches.", ins_r, "insn"), + ]) + + def Far() -> MetricGroup: + br = Event("ex_ret_brn_far") + br_r = d_ratio(br, interval_sec) + ins_r = d_ratio(ins, br) + return MetricGroup("br_far", [ + Metric("br_far_retired", "Retired far control transfers per second.", + br_r, "insn/s"), + Metric( + "br_far_insn_between_branches", + "The number of instructions divided by the number of far branches.", + ins_r, "insn"), + ]) + + return MetricGroup("br", [Total(), Taken(), Conditional(), Fused(), Far()], + description="breakdown of retired branch instructions") + + def AmdUpc() -> Metric: ops = Event("ex_ret_ops", "ex_ret_cops") upc = d_ratio(ops, smt_cycles)
@@ -58,6 +161,7 @@ def Rapl() -> MetricGroup: description="Processor socket power consumption estimates") all_metrics = MetricGroup("", [ + AmdBr(), AmdUpc(), Idle(), Rapl(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 5/12 perf jevents: Add software prefetch (swpf) metric group for AMD
In reply to Ian RogersAdd metrics that give the utility of software prefetches on zen2, zen3 and zen4.
--- tools/perf/pmu-events/amd_metrics.py | 95 ++++++++++++++++++++++++++++ 1 file changed, 95 insertions(+)
tools/perf/pmu-events/amd_metrics.py +95 -0
@@ -125,6 +125,100 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdSwpf() -> Optional[MetricGroup]: + """Returns a MetricGroup representing AMD software prefetch metrics.""" + if zen_model <= 1: + return None + + swp_ld = Event("ls_dispatch.ld_dispatch") + swp_t0 = Event("ls_pref_instr_disp.prefetch") + swp_w = Event("ls_pref_instr_disp.prefetch_w") # Missing on Zen1 + swp_nt = Event("ls_pref_instr_disp.prefetch_nta") + swp_mab = Event("ls_inef_sw_pref.mab_mch_cnt") + swp_l2 = Event("ls_sw_pf_dc_fills.local_l2", + "ls_sw_pf_dc_fills.lcl_l2", + "ls_sw_pf_dc_fill.ls_mabresp_lcl_l2") + swp_lc = Event("ls_sw_pf_dc_fills.local_ccx", + "ls_sw_pf_dc_fills.int_cache", + "ls_sw_pf_dc_fill.ls_mabresp_lcl_cache") + swp_lm = Event("ls_sw_pf_dc_fills.dram_io_near", + "ls_sw_pf_dc_fills.mem_io_local", + "ls_sw_pf_dc_fill.ls_mabresp_lcl_dram") + swp_rc = Event("ls_sw_pf_dc_fills.far_cache", + "ls_sw_pf_dc_fills.ext_cache_remote", + "ls_sw_pf_dc_fill.ls_mabresp_rmt_cache") + swp_rm = Event("ls_sw_pf_dc_fills.dram_io_far", + "ls_sw_pf_dc_fills.mem_io_remote", + "ls_sw_pf_dc_fill.ls_mabresp_rmt_dram") + + # All the swpf that were satisfied beyond L1D are good. + all_pf = swp_t0 + swp_w + swp_nt + good_pf = swp_l2 + swp_lc + swp_lm + swp_rc + swp_rm + bad_pf = max(all_pf - good_pf, 0) + + loc_pf = swp_l2 + swp_lc + swp_lm + rem_pf = swp_rc + swp_rm + + req_pend = max(0, bad_pf - swp_mab) + + r1 = d_ratio(ins, all_pf) + r2 = d_ratio(swp_ld, all_pf) + r3 = d_ratio(swp_t0, interval_sec) + r4 = d_ratio(swp_w, interval_sec) + r5 = d_ratio(swp_nt, interval_sec) + overview = MetricGroup("swpf_overview", [ + Metric("swpf_ov_insn_bt_swpf", "Insn between SWPF", r1, "insns"), + Metric("swpf_ov_loads_bt_swpf", "Loads between SWPF", r2, "loads"), + Metric("swpf_ov_rate_prefetch_t0_t1_t2", "Rate prefetch TO_T1_T2", r3, + "insns/sec"), + Metric("swpf_ov_rate_prefetch_w", "Rate prefetch W", r4, "insns/sec"), + Metric("swpf_ov_rate_preftech_nta", "Rate prefetch NTA", r5, "insns/sec"), + ]) + + r1 = d_ratio(swp_mab, all_pf) + r2 = d_ratio(req_pend, all_pf) + usefulness_bad = MetricGroup("swpf_usefulness_bad", [ + Metric("swpf_use_bad_hit_l1", "Usefulness bad hit L1", r1, "100%"), + Metric("swpf_use_bad_req_pend", "Usefulness bad req pending", r2, "100%"), + ]) + + r1 = d_ratio(good_pf, all_pf) + usefulness_good = MetricGroup("swpf_usefulness_good", [ + Metric("swpf_use_good_other_src", "Usefulness good other src", r1, + "100%"), + ]) + + usefulness = MetricGroup("swpf_usefulness", [ + usefulness_bad, + usefulness_good, + ]) + + r1 = d_ratio(swp_l2, good_pf) + r2 = d_ratio(swp_lc, good_pf) + r3 = d_ratio(swp_lm, good_pf) + data_src_local = MetricGroup("swpf_data_src_local", [ + Metric("swpf_data_src_local_l2", "Data source local l2", r1, "100%"), + Metric("swpf_data_src_local_ccx_l3_loc_ccx", + "Data source local ccx l3 loc ccx", r2, "100%"), + Metric("swpf_data_src_local_memory_or_io", + "Data source local memory or IO", r3, "100%"), + ]) + + r1 = d_ratio(swp_rc, good_pf) + r2 = d_ratio(swp_rm, good_pf) + data_src_remote = MetricGroup("swpf_data_src_remote", [ + Metric("swpf_data_src_remote_cache", "Data source remote cache", r1, + "100%"), + Metric("swpf_data_src_remote_memory_or_io", + "Data source remote memory or IO", r2, "100%"), + ]) + + data_src = MetricGroup("swpf_data_src", [data_src_local, data_src_remote]) + + return MetricGroup("swpf", [overview, usefulness, data_src], + description="Sofware prefetch breakdown (CCX L3 = L3 of current thread, Loc CCX = CCX cache on some socket)") + + def AmdUpc() -> Metric: ops = Event("ex_ret_ops", "ex_ret_cops") upc = d_ratio(ops, smt_cycles)
@@ -162,6 +256,7 @@ def Rapl() -> MetricGroup: all_metrics = MetricGroup("", [ AmdBr(), + AmdSwpf(), AmdUpc(), Idle(), Rapl(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 6/12 perf jevents: Add hardware prefetch (hwpf) metric group for AMD
In reply to Ian RogersAdd metrics that give the utility of hardware prefetches on zen2, zen3 and zen4.
--- tools/perf/pmu-events/amd_metrics.py | 61 ++++++++++++++++++++++++++++ 1 file changed, 61 insertions(+)
tools/perf/pmu-events/amd_metrics.py +61 -0
@@ -125,6 +125,66 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdHwpf(): + """Returns a MetricGroup representing AMD hardware prefetch metrics.""" + if zen_model <= 1: + return None + + hwp_ld = Event("ls_dispatch.ld_dispatch") + hwp_l2 = Event("ls_hw_pf_dc_fills.local_l2", + "ls_hw_pf_dc_fills.lcl_l2", + "ls_hw_pf_dc_fill.ls_mabresp_lcl_l2") + hwp_lc = Event("ls_hw_pf_dc_fills.local_ccx", + "ls_hw_pf_dc_fills.int_cache", + "ls_hw_pf_dc_fill.ls_mabresp_lcl_cache") + hwp_lm = Event("ls_hw_pf_dc_fills.dram_io_near", + "ls_hw_pf_dc_fills.mem_io_local", + "ls_hw_pf_dc_fill.ls_mabresp_lcl_dram") + hwp_rc = Event("ls_hw_pf_dc_fills.far_cache", + "ls_hw_pf_dc_fills.ext_cache_remote", + "ls_hw_pf_dc_fill.ls_mabresp_rmt_cache") + hwp_rm = Event("ls_hw_pf_dc_fills.dram_io_far", + "ls_hw_pf_dc_fills.mem_io_remote", + "ls_hw_pf_dc_fill.ls_mabresp_rmt_dram") + + loc_pf = hwp_l2 + hwp_lc + hwp_lm + rem_pf = hwp_rc + hwp_rm + all_pf = loc_pf + rem_pf + + r1 = d_ratio(ins, all_pf) + r2 = d_ratio(hwp_ld, all_pf) + r3 = d_ratio(all_pf, interval_sec) + + overview = MetricGroup("hwpf_overview", [ + Metric("hwpf_ov_insn_bt_hwpf", "Insn between HWPF", r1, "insns"), + Metric("hwpf_ov_loads_bt_hwpf", "Loads between HWPF", r2, "loads"), + Metric("hwpf_ov_rate", "HWPF per second", r3, "hwpf/s"), + ]) + r1 = d_ratio(hwp_l2, all_pf) + r2 = d_ratio(hwp_lc, all_pf) + r3 = d_ratio(hwp_lm, all_pf) + data_src_local = MetricGroup("hwpf_data_src_local", [ + Metric("hwpf_data_src_local_l2", "Data source local l2", r1, "100%"), + Metric("hwpf_data_src_local_ccx_l3_loc_ccx", + "Data source local ccx l3 loc ccx", r2, "100%"), + Metric("hwpf_data_src_local_memory_or_io", + "Data source local memory or IO", r3, "100%"), + ]) + + r1 = d_ratio(hwp_rc, all_pf) + r2 = d_ratio(hwp_rm, all_pf) + data_src_remote = MetricGroup("hwpf_data_src_remote", [ + Metric("hwpf_data_src_remote_cache", "Data source remote cache", r1, + "100%"), + Metric("hwpf_data_src_remote_memory_or_io", + "Data source remote memory or IO", r2, "100%"), + ]) + + data_src = MetricGroup("hwpf_data_src", [data_src_local, data_src_remote]) + return MetricGroup("hwpf", [overview, data_src], + description="Hardware prefetch breakdown (CCX L3 = L3 of current thread, Loc CCX = CCX cache on some socket)") + + def AmdSwpf() -> Optional[MetricGroup]: """Returns a MetricGroup representing AMD software prefetch metrics.""" if zen_model <= 1:
@@ -256,6 +316,7 @@ def Rapl() -> MetricGroup: all_metrics = MetricGroup("", [ AmdBr(), + AmdHwpf(), AmdSwpf(), AmdUpc(), Idle(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 7/12 perf jevents: Add itlb metric group for AMD
In reply to Ian RogersAdd metrics that give an overview and details of the l1 itlb (zen1, zen2, zen3) and l2 itlb (all zens).
--- tools/perf/pmu-events/amd_metrics.py | 47 ++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+)
tools/perf/pmu-events/amd_metrics.py +47 -0
@@ -125,6 +125,52 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdItlb(): + l2h = Event("bp_l1_tlb_miss_l2_tlb_hit", "bp_l1_tlb_miss_l2_hit") + l2m = Event("l2_itlb_misses") + l2r = l2h + l2m + + itlb_l1_mg = None + l1m = l2r + if zen_model <= 3: + l1r = Event("ic_fw32") + l1h = max(l1r - l1m, 0) + itlb_l1_mg = MetricGroup("itlb_l1", [ + Metric("itlb_l1_hits", + "L1 ITLB hits as a perecentage of L1 ITLB accesses.", + d_ratio(l1h, l1h + l1m), "100%"), + Metric("itlb_l1_miss", + "L1 ITLB misses as a perecentage of L1 ITLB accesses.", + d_ratio(l1m, l1h + l1m), "100%"), + Metric("itlb_l1_reqs", + "The number of 32B fetch windows transferred from IC pipe to DE " + "instruction decoder per second.", d_ratio(l1r, interval_sec), + "windows/sec"), + ]) + + return MetricGroup("itlb", [ + MetricGroup("itlb_ov", [ + Metric("itlb_ov_insn_bt_l1_miss", + "Number of instructions between l1 misses", d_ratio( + ins, l1m), "insns"), + Metric("itlb_ov_insn_bt_l2_miss", + "Number of instructions between l2 misses", d_ratio( + ins, l2m), "insns"), + ]), + itlb_l1_mg, + MetricGroup("itlb_l2", [ + Metric("itlb_l2_hits", + "L2 ITLB hits as a percentage of all L2 ITLB accesses.", + d_ratio(l2h, l2r), "100%"), + Metric("itlb_l2_miss", + "L2 ITLB misses as a percentage of all L2 ITLB accesses.", + d_ratio(l2m, l2r), "100%"), + Metric("itlb_l2_reqs", "ITLB accesses per second.", + d_ratio(l2r, interval_sec), "accesses/sec"), + ]), + ], description="Instruction TLB breakdown") + + def AmdHwpf(): """Returns a MetricGroup representing AMD hardware prefetch metrics.""" if zen_model <= 1:
@@ -316,6 +362,7 @@ def Rapl() -> MetricGroup: all_metrics = MetricGroup("", [ AmdBr(), + AmdItlb(), AmdHwpf(), AmdSwpf(), AmdUpc(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 8/12 perf jevents: Add dtlb metric group for AMD
In reply to Ian RogersAdd metrics that give an overview and details of the dtlb (zen1, zen2, zen3).
--- tools/perf/pmu-events/amd_metrics.py | 108 +++++++++++++++++++++++++++ 1 file changed, 108 insertions(+)
tools/perf/pmu-events/amd_metrics.py +108 -0
@@ -125,6 +125,113 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdDtlb() -> Optional[MetricGroup]: + if zen_model >= 4: + return None + + d_dat = Event("ls_dc_accesses") if zen_model <= 3 else None + d_h4k = Event("ls_l1_d_tlb_miss.tlb_reload_4k_l2_hit") + d_hcoal = Event("ls_l1_d_tlb_miss.tlb_reload_coalesced_page_hit") if zen_model >= 2 else 0 + d_h2m = Event("ls_l1_d_tlb_miss.tlb_reload_2m_l2_hit") + d_h1g = Event("ls_l1_d_tlb_miss.tlb_reload_1g_l2_hit") + + d_m4k = Event("ls_l1_d_tlb_miss.tlb_reload_4k_l2_miss") + d_mcoal = Event("ls_l1_d_tlb_miss.tlb_reload_coalesced_page_miss") if zen_model >= 2 else 0 + d_m2m = Event("ls_l1_d_tlb_miss.tlb_reload_2m_l2_miss") + d_m1g = Event("ls_l1_d_tlb_miss.tlb_reload_1g_l2_miss") + + d_w0 = Event("ls_tablewalker.dc_type0") if zen_model <= 3 else None + d_w1 = Event("ls_tablewalker.dc_type1") if zen_model <= 3 else None + walks = d_w0 + d_w1 + walks_r = d_ratio(walks, interval_sec) + ins_w = d_ratio(ins, walks) + l1 = d_dat + l1_r = d_ratio(l1, interval_sec) + l2_hits = d_h4k + d_hcoal + d_h2m + d_h1g + l2_miss = d_m4k + d_mcoal + d_m2m + d_m1g + l2_r = d_ratio(l2_hits + l2_miss, interval_sec) + l1_miss = l2_hits + l2_miss + walks + l1_hits = max(l1 - l1_miss, 0) + ins_l = d_ratio(ins, l1_miss) + + return MetricGroup("dtlb", [ + MetricGroup("dtlb_ov", [ + Metric("dtlb_ov_insn_bt_l1_miss", + "DTLB overview: instructions between l1 misses.", ins_l, + "insns"), + Metric("dtlb_ov_insn_bt_walks", + "DTLB overview: instructions between dtlb page table walks.", + ins_w, "insns"), + ]), + MetricGroup("dtlb_l1", [ + Metric("dtlb_l1_hits", + "DTLB L1 hits as percentage of all DTLB L1 accesses.", + d_ratio(l1_hits, l1), "100%"), + Metric("dtlb_l1_miss", + "DTLB L1 misses as percentage of all DTLB L1 accesses.", + d_ratio(l1_miss, l1), "100%"), + Metric("dtlb_l1_reqs", "DTLB L1 accesses per second.", l1_r, + "insns/s"), + ]), + MetricGroup("dtlb_l2", [ + Metric("dtlb_l2_hits", + "DTLB L2 hits as percentage of all DTLB L2 accesses.", + d_ratio(l2_hits, l2_hits + l2_miss), "100%"), + Metric("dtlb_l2_miss", + "DTLB L2 misses as percentage of all DTLB L2 accesses.", + d_ratio(l2_miss, l2_hits + l2_miss), "100%"), + Metric("dtlb_l2_reqs", "DTLB L2 accesses per second.", l2_r, + "insns/s"), + MetricGroup("dtlb_l2_4kb", [ + Metric( + "dtlb_l2_4kb_hits", + "DTLB L2 4kb page size hits as percentage of all DTLB L2 4kb " + "accesses.", d_ratio(d_h4k, d_h4k + d_m4k), "100%"), + Metric( + "dtlb_l2_4kb_miss", + "DTLB L2 4kb page size misses as percentage of all DTLB L2 4kb" + "accesses.", d_ratio(d_m4k, d_h4k + d_m4k), "100%") + ]), + MetricGroup("dtlb_l2_coalesced", [ + Metric( + "dtlb_l2_coal_hits", + "DTLB L2 coalesced page (16kb) hits as percentage of all DTLB " + "L2 coalesced accesses.", d_ratio(d_hcoal, + d_hcoal + d_mcoal), "100%"), + Metric( + "dtlb_l2_coal_miss", + "DTLB L2 coalesced page (16kb) misses as percentage of all " + "DTLB L2 coalesced accesses.", + d_ratio(d_mcoal, d_hcoal + d_mcoal), "100%") + ]), + MetricGroup("dtlb_l2_2mb", [ + Metric( + "dtlb_l2_2mb_hits", + "DTLB L2 2mb page size hits as percentage of all DTLB L2 2mb " + "accesses.", d_ratio(d_h2m, d_h2m + d_m2m), "100%"), + Metric( + "dtlb_l2_2mb_miss", + "DTLB L2 2mb page size misses as percentage of all DTLB L2 " + "accesses.", d_ratio(d_m2m, d_h2m + d_m2m), "100%") + ]), + MetricGroup("dtlb_l2_1g", [ + Metric( + "dtlb_l2_1g_hits", + "DTLB L2 1gb page size hits as percentage of all DTLB L2 1gb " + "accesses.", d_ratio(d_h1g, d_h1g + d_m1g), "100%"), + Metric( + "dtlb_l2_1g_miss", + "DTLB L2 1gb page size misses as percentage of all DTLB L2 " + "1gb accesses.", d_ratio(d_m1g, d_h1g + d_m1g), "100%") + ]), + ]), + MetricGroup("dtlb_walks", [ + Metric("dtlb_walks_reqs", "DTLB page table walks per second.", + walks_r, "walks/s"), + ]), + ], description="Data TLB metrics") + + def AmdItlb(): l2h = Event("bp_l1_tlb_miss_l2_tlb_hit", "bp_l1_tlb_miss_l2_hit") l2m = Event("l2_itlb_misses")
@@ -362,6 +469,7 @@ def Rapl() -> MetricGroup: all_metrics = MetricGroup("", [ AmdBr(), + AmdDtlb(), AmdItlb(), AmdHwpf(), AmdSwpf(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 9/12 perf jevents: Add uncore l3 metric group for AMD
In reply to Ian RogersMetrics use the amd_l3 PMU for access/miss/hit information.
--- tools/perf/pmu-events/amd_metrics.py | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
tools/perf/pmu-events/amd_metrics.py +18 -0
@@ -467,6 +467,23 @@ def Rapl() -> MetricGroup: return MetricGroup("cpu_power", metrics, description="Processor socket power consumption estimates") +def UncoreL3(): + acc = Event("l3_lookup_state.all_coherent_accesses_to_l3", + "l3_lookup_state.all_l3_req_typs") + miss = Event("l3_lookup_state.l3_miss", + "l3_comb_clstr_state.request_miss") + acc = max(acc, miss) + hits = acc - miss + + return MetricGroup("l3", [ + Metric("l3_accesses", "L3 victim cache accesses", + d_ratio(acc, interval_sec), "accesses/sec"), + Metric("l3_hits", "L3 victim cache hit rate", d_ratio(hits, acc), "100%"), + Metric("l3_miss", "L3 victim cache miss rate", d_ratio(miss, acc), + "100%"), + ], description="L3 cache breakdown per CCX") + + all_metrics = MetricGroup("", [ AmdBr(), AmdDtlb(),
@@ -476,6 +493,7 @@ all_metrics = MetricGroup("", [ AmdUpc(), Idle(), Rapl(), + UncoreL3(), ]) if args.metricgroups:
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 10/12 perf jevents: Add load store breakdown metrics ldst for AMD
In reply to Ian RogersGive breakdown of number of instructions. Use the counter mask (cmask) to show the number of cycles taken to retire the instructions.
--- tools/perf/pmu-events/amd_metrics.py | 75 ++++++++++++++++++++++++++++ 1 file changed, 75 insertions(+)
tools/perf/pmu-events/amd_metrics.py +75 -0
@@ -278,6 +278,80 @@ def AmdItlb(): ], description="Instruction TLB breakdown") +def AmdLdSt() -> MetricGroup: + ldst_ld = Event("ls_dispatch.ld_dispatch") + ldst_st = Event("ls_dispatch.store_dispatch") + ldst_ldc1 = Event(f"{ldst_ld}/cmask=1/") + ldst_stc1 = Event(f"{ldst_st}/cmask=1/") + ldst_ldc2 = Event(f"{ldst_ld}/cmask=2/") + ldst_stc2 = Event(f"{ldst_st}/cmask=2/") + ldst_ldc3 = Event(f"{ldst_ld}/cmask=3/") + ldst_stc3 = Event(f"{ldst_st}/cmask=3/") + ldst_cyc = Event("ls_not_halted_cyc") + + ld_rate = d_ratio(ldst_ld, interval_sec) + st_rate = d_ratio(ldst_st, interval_sec) + + ld_v1 = max(ldst_ldc1 - ldst_ldc2, 0) + ld_v2 = max(ldst_ldc2 - ldst_ldc3, 0) + ld_v3 = ldst_ldc3 + + st_v1 = max(ldst_stc1 - ldst_stc2, 0) + st_v2 = max(ldst_stc2 - ldst_stc3, 0) + st_v3 = ldst_stc3 + + return MetricGroup("ldst", [ + MetricGroup("ldst_total", [ + Metric("ldst_total_ld", "Number of loads dispatched per second.", + ld_rate, "insns/sec"), + Metric("ldst_total_st", "Number of stores dispatched per second.", + st_rate, "insns/sec"), + ]), + MetricGroup("ldst_percent_insn", [ + Metric("ldst_percent_insn_ld", + "Load instructions as a percentage of all instructions.", + d_ratio(ldst_ld, ins), "100%"), + Metric("ldst_percent_insn_st", + "Store instructions as a percentage of all instructions.", + d_ratio(ldst_st, ins), "100%"), + ]), + MetricGroup("ldst_ret_loads_per_cycle", [ + Metric( + "ldst_ret_loads_per_cycle_1", + "Load instructions retiring in 1 cycle as a percentage of all " + "unhalted cycles.", d_ratio(ld_v1, ldst_cyc), "100%"), + Metric( + "ldst_ret_loads_per_cycle_2", + "Load instructions retiring in 2 cycles as a percentage of all " + "unhalted cycles.", d_ratio(ld_v2, ldst_cyc), "100%"), + Metric( + "ldst_ret_loads_per_cycle_3", + "Load instructions retiring in 3 or more cycles as a percentage" + "of all unhalted cycles.", d_ratio(ld_v3, ldst_cyc), "100%"), + ]), + MetricGroup("ldst_ret_stores_per_cycle", [ + Metric( + "ldst_ret_stores_per_cycle_1", + "Store instructions retiring in 1 cycle as a percentage of all " + "unhalted cycles.", d_ratio(st_v1, ldst_cyc), "100%"), + Metric( + "ldst_ret_stores_per_cycle_2", + "Store instructions retiring in 2 cycles as a percentage of all " + "unhalted cycles.", d_ratio(st_v2, ldst_cyc), "100%"), + Metric( + "ldst_ret_stores_per_cycle_3", + "Store instructions retiring in 3 or more cycles as a percentage" + "of all unhalted cycles.", d_ratio(st_v3, ldst_cyc), "100%"), + ]), + MetricGroup("ldst_insn_bt", [ + Metric("ldst_insn_bt_ld", "Number of instructions between loads.", + d_ratio(ins, ldst_ld), "insns"), + Metric("ldst_insn_bt_st", "Number of instructions between stores.", + d_ratio(ins, ldst_st), "insns"), + ]) + ], description="Breakdown of load/store instructions") + + def AmdHwpf(): """Returns a MetricGroup representing AMD hardware prefetch metrics.""" if zen_model <= 1:
@@ -488,6 +562,7 @@ all_metrics = MetricGroup("", [ AmdBr(), AmdDtlb(), AmdItlb(), + AmdLdSt(), AmdHwpf(), AmdSwpf(), AmdUpc(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 11/12 perf jevents: Add ILP metrics for AMD
In reply to Ian RogersUse the counter mask (cmask) to see how many cycles an instruction takes to retire. Present as a set of ILP metrics.
--- tools/perf/pmu-events/amd_metrics.py | 30 ++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+)
tools/perf/pmu-events/amd_metrics.py +30 -0
@@ -125,6 +125,35 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdIlp() -> MetricGroup: + tsc = Event("msr/tsc/") + c0 = Event("msr/mperf/") + low = tsc - c0 + inst_ret = Event("ex_ret_instr") + inst_ret_c = [Event(f"{inst_ret.name}/cmask={x}/") for x in range(1, 6)] + ilp = [d_ratio(max(inst_ret_c[x] - inst_ret_c[x + 1], 0), cycles) for x in range(0, 4)] + ilp.append(d_ratio(inst_ret_c[4], cycles)) + ilp0 = 1 + for x in ilp: + ilp0 -= x + return MetricGroup("ilp", [ + Metric("ilp_idle", "Lower power cycles as a percentage of all cycles", + d_ratio(low, tsc), "100%"), + Metric("ilp_inst_ret_0", "Instructions retired in 0 cycles as a percentage of all cycles", + ilp0, "100%"), + Metric("ilp_inst_ret_1", "Instructions retired in 1 cycles as a percentage of all cycles", + ilp[0], "100%"), + Metric("ilp_inst_ret_2", "Instructions retired in 2 cycles as a percentage of all cycles", + ilp[1], "100%"), + Metric("ilp_inst_ret_3", "Instructions retired in 3 cycles as a percentage of all cycles", + ilp[2], "100%"), + Metric("ilp_inst_ret_4", "Instructions retired in 4 cycles as a percentage of all cycles", + ilp[3], "100%"), + Metric("ilp_inst_ret_5", "Instructions retired in 5 or more cycles as a percentage of all cycles", + ilp[4], "100%"), + ]) + + def AmdDtlb() -> Optional[MetricGroup]: if zen_model >= 4: return None
@@ -560,6 +589,7 @@ def UncoreL3(): all_metrics = MetricGroup("", [ AmdBr(), + AmdIlp(), AmdDtlb(), AmdItlb(), AmdLdSt(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
Ian Rogers
reviewerPATCH v2 12/12 perf jevents: Add context switch metrics for AMD
In reply to Ian RogersMetrics break down context switches for different kinds of instruction.
--- tools/perf/pmu-events/amd_metrics.py | 32 ++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+)
tools/perf/pmu-events/amd_metrics.py +32 -0
@@ -125,6 +125,37 @@ def AmdBr(): description="breakdown of retired branch instructions") +def AmdCtxSw() -> MetricGroup: + cs = Event("context\-switches") + metrics = [ + Metric("cs_rate", "Context switches per second", d_ratio(cs, interval_sec), "ctxsw/s") + ] + + ev = Event("instructions") + metrics.append(Metric("cs_instr", "Instructions per context switch", + d_ratio(ev, cs), "instr/cs")) + + ev = Event("cycles") + metrics.append(Metric("cs_cycles", "Cycles per context switch", + d_ratio(ev, cs), "cycles/cs")) + + ev = Event("ls_dispatch.ld_dispatch") + metrics.append(Metric("cs_loads", "Loads per context switch", + d_ratio(ev, cs), "loads/cs")) + + ev = Event("ls_dispatch.store_dispatch") + metrics.append(Metric("cs_stores", "Stores per context switch", + d_ratio(ev, cs), "stores/cs")) + + ev = Event("ex_ret_brn_tkn") + metrics.append(Metric("cs_br_taken", "Branches taken per context switch", + d_ratio(ev, cs), "br_taken/cs")) + + return MetricGroup("cs", metrics, + description = ("Number of context switches per second, instructions " + "retired & core cycles between context switches")) + + def AmdIlp() -> MetricGroup: tsc = Event("msr/tsc/") c0 = Event("msr/mperf/")
@@ -589,6 +620,7 @@ def UncoreL3(): all_metrics = MetricGroup("", [ AmdBr(), + AmdCtxSw(), AmdIlp(), AmdDtlb(), AmdItlb(),
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog