#!/bin/bash # # icache - measure instruction cache usage. # Uses Linux perf and PMCs. # # FROM: https://github.com/brendangregg/pmc-cloud-tools # # USAGE: icache {-C CPU | -p PID | -c CMD} [interval [duration]] # # Columns: # # - INS: Instructions x 1000 # - L1IMISS: Level 1 instruction cache misses x 1000 # - L1I%: Level 1 instruction cache hit ratio percent (L1MISS/INS) # - L2IHIT: Level 2 instruction cache hits x 1000 # - L2IMISS: Level 2 instruction cache misses x 1000 # - L2I%: Level 2 instruction cache hit ratio percent # - LLCREF: Last Level Cache references x 1000 # - LLCMISS: Last Level Cache misses x 1000 # - LLC%: Last Level Cache hit ratio percent # # Copyright 2020 Netflix, Inc. # Licensed under the Apache License, Version 2.0 (the "License") # # 24-Jan-2020 Brendan Gregg Created this. function usage { cat <<-END >&2 USAGE: icache {-C CPU | -p PID | -c CMD} [interval [duration]] -C CPU # measure this CPU only -p PID # measure this PID only -c 'CMD' # measure this command only (quote it) interval # output interval in secs (default 1) duration # total seconds (default infinityish) eg, icache # show stats across all CPUs icache 5 # show stats every 5 seconds icache -C 0 # measure CPU 0 only icache -p 181 # measure PID 181 only icache -c 'cksum /boot/*' # measure run and measure this cmd END exit } opt_cpu=0; opt_pid=0; opt_cmd=0; cpu=""; pid=""; cmd="" while getopts C:p:c:h opt do case $opt in C) opt_cpu=1; cpu=$OPTARG ;; p) opt_pid=1; pid=$OPTARG ;; c) opt_cmd=1; cmd=$OPTARG ;; h|?) usage ;; esac done shift $(( $OPTIND - 1 )) if (( opt_cpu + opt_pid + opt_cmd > 1 )); then echo >&2 "ERROR: pick one of -C, -p, -c" usage fi secs=${1:-1} # default 1 second duration=${2:-999999999} # default semi-infinite seconds hlines=25 # lines to repeat header target=-a (( opt_cpu )) && target="-C $cpu sleep $duration" (( opt_pid )) && target="-p $pid sleep $duration" (( opt_cmd )) && target="$cmd" if (( opt_pid )); then if [ ! -d /proc/$pid ]; then echo >&2 "ERROR: Can't find PID $pid. Exiting." exit fi fi # note that instructions is last on purpose, it triggers output # cycles are twice as a workaround for an issue # the r4f2e and r412e counters are from the architectural set, so should be stable echo "All counter columns are x 1000" perf stat -e instructions,L1-icache-load-misses,r2424,r4424,cache-references,cache-misses \ -I $(( secs * 1000 )) $target 2>&1 | awk -v hlines=$hlines ' BEGIN { htxt = sprintf("%-10s %10s %5s %9s %9s %5s %8s %8s %5s", "INS", "L1IMISS", "L1I%", "L2IHIT", "L2IMISS", "L2I%", "LLCREF", "LLCMISS", "LLC%"); print htxt header = hlines } /invalid/ { print $0 } # unsupported event { gsub(/,/, ""); } $3 == "instructions" { ins = $2 } $3 == "L1-icache-load-misses" { l1im = $2 } $3 == "r2424" { l2im = $2 } # L2_RQSTS.CODE_RD_MISS $3 == "r4424" { l2ih = $2 } # L2_RQSTS.CODE_RD_HIT $3 == "cache-references" { llcr = $2 } $3 == "cache-misses" { # last one, trigger output llcm = $2 if (--header == 0) { print htxt header = hlines } # if we are missing refs, then also reset hits to zero, to avoid a bogus ratio: if (ins == 0) { ins = 1; l1im = 0 } if (l2im == 0) { l2im = 1; l2ih = 0 } if (llcr == 0) { llcr = 1; llcm = 0 } l1iratio = 100 * (ins - l1im) / ins l2iratio = 100 * l2ih / (l2ih + l2im) llcratio = 100 * (llcr - llcm) / llcr printf("%-10d %10d %5.2f %9d %9d %5.2f %8d %8d %5.2f\n", ins / 1000, l1im / 1000, l1iratio, l2ih / 1000, l2im / 1000, l2iratio, llcr / 1000, llcm / 1000, llcratio); } '