Merge pull request #68 from intel/prep140

changes for version 1.4.0

Author: harp-intel

README.md

@@ -41,6 +41,16 @@ sudo ./perf-collect --timeout 10
 ./perf-postprocess
 ```
 
+## Running perf-collect as a non-root user
+As seen in the examples above, `sudo` is the standard approach to running perf-collect with elevated privileges. If `sudo` is not possible and running as the root user is not possible, then a user may request the following changes be made to the system by an administrator:
+- sysctl -w kernel.perf_event_paranoid=0
+- sysctl -w kernel.nmi_watchdog=0
+- write '125' to all perf_event_mux_interval_ms files found under /sys/devices/*.
+
+`for i in $(find /sys/devices -name perf_event_mux_interval_ms); do echo 125 > $i; done`
+
+Recommend returning these settings to their prior values when analysis with PerfSpect is complete.
+
 ## Output
 
 perf-collect outputs:

_version.txt

@@ -1 +1 @@
-1.3.11
+1.4.0

events/icx_nofixedtma.txt

@@ -0,0 +1,148 @@
+###########################################################################################################
+# Copyright (C) 2021-2023 Intel Corporation
+# SPDX-License-Identifier: BSD-3-Clause
+###########################################################################################################
+
+# Icelake event list for platforms that don't have support for the fixed counter TMA events, e.g., some AWS
+# VMs.
+# Note that there are no more than 10 events per group. On these same platforms, the cpu-cycles fixed
+# counter is not supported so a general purpose counter will be used.
+
+cpu/event=0x51,umask=0x01,period=100003,name='L1D.REPLACEMENT'/,
+cpu/event=0xd1,umask=0x01,period=1000003,name='MEM_LOAD_RETIRED.L1_HIT'/,
+cpu/event=0x24,umask=0xe4,period=200003,name='L2_RQSTS.ALL_CODE_RD'/,
+cpu/event=0xc3,umask=0x01,cmask=0x01,edge=0x01,period=100003,name='MACHINE_CLEARS.COUNT'/,
+cpu/event=0xc5,umask=0x00,period=50021,name='BR_MISP_RETIRED.ALL_BRANCHES'/,
+cpu/event=0xf1,umask=0x1f,period=100003,name='L2_LINES_IN.ALL'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0xd1,umask=0x10,period=100021,name='MEM_LOAD_RETIRED.L2_MISS'/,
+cpu/event=0x79,umask=0x08,cmask=0x00,period=2000003,name='IDQ.DSB_UOPS'/,
+cpu/event=0xa8,umask=0x01,cmask=0x00,period=2000003,name='LSD.UOPS'/,
+cpu/event=0x48,umask=0x02,period=1000003,name='L1D_PEND_MISS.FB_FULL_PERIODS'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+# events for TMA metrics without fixed counter support (group 1)
+cpu/event=0x9c,umask=0x01,name='IDQ_UOPS_NOT_DELIVERED.CORE'/,
+cpu/event=0xa4,umask=0x01,name='TOPDOWN.SLOTS_P'/,
+cpu/event=0x80,umask=0x04,name='ICACHE_DATA.STALLS'/,
+cpu/event=0x83,umask=0x04,name='ICACHE_TAG.STALLS'/,
+cpu/event=0x79,umask=0x30,name='IDQ.MS_SWITCHES'/,
+cpu/event=0x87,umask=0x01,name='DECODE.LCP'/,
+cpu/event=0x0d,umask=0x10,period=1000003,name='INT_MISC.UOP_DROPPING'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+# events for TMA metrics without fixed counter support (group 2)
+cpu/event=0xab,umask=0x02,name='DSB2MITE_SWITCHES.PENALTY_CYCLES'/,
+cpu/event=0xa4,umask=0x02,name='TOPDOWN.BACKEND_BOUND_SLOTS'/,
+cpu/event=0x0D,umask=0x01,name='INT_MISC.CLEARS_COUNT'/,
+cpu/event=0xc2,umask=0x02,name='UOPS_RETIRED.SLOTS'/,
+cpu/event=0xd0,umask=0x83,name='MEM_INST_RETIRED.ANY'/,
+cpu/event=0xc4,umask=0x00,name='BR_INST_RETIRED.ALL_BRANCHES'/,
+cpu/event=0x9c,umask=0x01,cmask=0x05,period=1000003,name='IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0x24,umask=0x24,period=200003,name='L2_RQSTS.CODE_RD_MISS'/,
+cpu/event=0xa3,umask=0x0C,cmask=0x0C,period=1000003,name='CYCLE_ACTIVITY.STALLS_L1D_MISS'/,
+cpu/event=0xa3,umask=0x14,cmask=0x14,period=2000003,name='CYCLE_ACTIVITY.STALLS_MEM_ANY'/,
+cpu/event=0xa6,umask=0x40,cmask=0x02,period=1000003,name='EXE_ACTIVITY.BOUND_ON_STORES'/,
+cpu/event=0xa3,umask=0x04,cmask=0x04,period=1000003,name='CYCLE_ACTIVITY.STALLS_TOTAL'/,
+cpu/event=0xa6,umask=0x02,period=2000003,name='EXE_ACTIVITY.1_PORTS_UTIL'/,
+cpu/event=0xa6,umask=0x04,period=2000003,name='EXE_ACTIVITY.2_PORTS_UTIL'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0xd0,umask=0x21,cmask=0x00,period=100007,name='MEM_INST_RETIRED.LOCK_LOADS'/,
+cpu/event=0xd1,umask=0x02,period=200003,name='MEM_LOAD_RETIRED.L2_HIT'/,
+cpu/event=0xd1,umask=0x40,period=100007,name='MEM_LOAD_RETIRED.FB_HIT'/,
+cpu/event=0xd1,umask=0x08,period=200003,name='MEM_LOAD_RETIRED.L1_MISS'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0xa3,umask=0x05,cmask=0x05,period=1000003,name='CYCLE_ACTIVITY.STALLS_L2_MISS'/,
+cpu/event=0xa3,umask=0x06,cmask=0x06,period=1000003,name='CYCLE_ACTIVITY.STALLS_L3_MISS'/,
+cpu/event=0xa3,umask=0x0c,cmask=0x0c,period=1000003,name='CYCLE_ACTIVITY.STALLS_L1D_MISS'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0x79,umask=0x04,cmask=0x01,period=2000003,name='IDQ.MITE_CYCLES_ANY'/,
+cpu/event=0x79,umask=0x04,cmask=0x05,period=2000003,name='IDQ.MITE_CYCLES_OK'/,
+cpu/event=0x79,umask=0x08,cmask=0x01,period=2000003,name='IDQ.DSB_CYCLES_ANY'/,
+cpu/event=0x79,umask=0x08,cmask=0x05,period=2000003,name='IDQ.DSB_CYCLES_OK'/,
+cpu/event=0xec,umask=0x02,period=2000003,name='CPU_CLK_UNHALTED.DISTRIBUTED'/,
+cpu/event=0x14,umask=0x09,cmask=0x01,period=1000003,name='ARITH.DIVIDER_ACTIVE'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0x79,umask=0x04,period=100003,name='IDQ.MITE_UOPS'/,
+cpu/event=0x79,umask=0x30,period=100003,name='IDQ.MS_UOPS'/,
+cpu/event=0x56,umask=0x01,period=100003,name='UOPS_DECODED.DEC0'/,
+cpu/event=0x56,umask=0x01,cmask=0x01,period=100003,name='UOPS_DECODED.DEC0:c1'/,
+cpu/event=0x0e,umask=0x01,period=2000003,name='UOPS_ISSUED.ANY'/,
+cpu-cycles:k,
+ref-cycles:k,
+instructions:k;
+
+# OCR
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x104000477,name='OCR.READS_TO_CORE.LOCAL_DRAM'/,
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x84002380,name='OCR.HWPF_L3.L3_MISS_LOCAL'/,
+cpu/event=0x85,umask=0x0e,period=100003,name='ITLB_MISSES.WALK_COMPLETED'/,
+cpu/event=0x08,umask=0x0e,period=100003,name='DTLB_LOAD_MISSES.WALK_COMPLETED'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x1030000477,name='OCR.READS_TO_CORE.REMOTE_CACHE.SNOOP_HITM'/,
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x830000477,name='OCR.READS_TO_CORE.REMOTE_CACHE.SNOOP_HIT_WITH_FWD'/,
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x730000477,name='OCR.READS_TO_CORE.REMOTE_DRAM'/,
+cpu/event=0xb7,umask=0x01,offcore_rsp=0x90002380,name='OCR.HWPF_L3.REMOTE'/,
+cpu/event=0x08,umask=0x04,period=100003,name='DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M'/,
+cpu/event=0x49,umask=0x0e,period=100003,name='DTLB_STORE_MISSES.WALK_COMPLETED'/,
+cpu-cycles,
+ref-cycles,
+instructions;
+
+# C6
+cstate_core/c6-residency/;
+cstate_pkg/c6-residency/;
+
+# UPI
+upi/event=0x2,umask=0xf,name='UNC_UPI_TxL_FLITS.ALL_DATA'/;
+
+# CHA
+cha/event=0x00,umask=0x00,name='UNC_CHA_CLOCKTICKS'/;
+
+cha/event=0x35,umask=0xC8177E01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_REMOTE'/,
+cha/event=0x35,umask=0xC816FE01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_LOCAL'/,
+cha/event=0x35,umask=0xC896FE01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_PREF_LOCAL'/,
+cha/event=0x35,umask=0xC8977E01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_PREF_REMOTE'/;
+
+cha/event=0x36,umask=0xc8177e01,name='UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD_REMOTE'/;
+cha/event=0x35,umask=0xc88ffe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_CRD_PREF'/,
+cha/event=0x35,umask=0xc80ffe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_CRD'/,
+cha/event=0x36,umask=0xC816FE01,name='UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD_LOCAL'/;
+
+cha/event=0x35,umask=0xccd7fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_LLCPREFDRD'/,
+cha/event=0x35,umask=0xc817fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD'/,
+cha/event=0x35,umask=0xc897fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_PREF'/,
+cha/event=0x36,umask=0xC817FE01,name='UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD'/;
+
+# memory read/writes
+imc/event=0x04,umask=0x0f,name='UNC_M_CAS_COUNT.RD'/,
+imc/event=0x04,umask=0x30,name='UNC_M_CAS_COUNT.WR'/;
+
+# power
+power/energy-pkg/,
+power/energy-ram/;