tips.txt (2881B)
1For a higher level overview, try: perf report --sort comm,dso 2Sample related events with: perf record -e '{cycles,instructions}:S' 3Compare performance results with: perf diff [<old file> <new file>] 4Boolean options have negative forms, e.g.: perf report --no-children 5Customize output of perf script with: perf script -F event,ip,sym 6Generate a script for your data: perf script -g <lang> 7Save output of perf stat using: perf stat record <target workload> 8Create an archive with symtabs to analyse on other machine: perf archive 9Search options using a keyword: perf report -h <keyword> 10Use parent filter to see specific call path: perf report -p <regex> 11List events using substring match: perf list <keyword> 12To see list of saved events and attributes: perf evlist -v 13Use --symfs <dir> if your symbol files are in non-standard locations 14To see callchains in a more compact form: perf report -g folded 15Show individual samples with: perf script 16Limit to show entries above 5% only: perf report --percent-limit 5 17Profiling branch (mis)predictions with: perf record -b / perf report 18To show assembler sample contexts use perf record -b / perf script -F +brstackinsn --xed 19Treat branches as callchains: perf report --branch-history 20To count events in every 1000 msec: perf stat -I 1000 21Print event counts in CSV format with: perf stat -x, 22If you have debuginfo enabled, try: perf report -s sym,srcline 23For memory address profiling, try: perf mem record / perf mem report 24For tracepoint events, try: perf report -s trace_fields 25To record callchains for each sample: perf record -g 26To record every process run by a user: perf record -u <user> 27Skip collecting build-id when recording: perf record -B 28To change sampling frequency to 100 Hz: perf record -F 100 29See assembly instructions with percentage: perf annotate <symbol> 30If you prefer Intel style assembly, try: perf annotate -M intel 31For hierarchical output, try: perf report --hierarchy 32Order by the overhead of source file name and line number: perf report -s srcline 33System-wide collection from all CPUs: perf record -a 34Show current config key-value pairs: perf config --list 35Show user configuration overrides: perf config --user --list 36To add Node.js USDT(User-Level Statically Defined Tracing): perf buildid-cache --add `which node` 37To report cacheline events from previous recording: perf c2c report 38To browse sample contexts use perf report --sample 10 and select in context menu 39To separate samples by time use perf report --sort time,overhead,sym 40To set sample time separation other than 100ms with --sort time use --time-quantum 41Add -I to perf record to sample register values, which will be visible in perf report sample context. 42To show IPC for sampling periods use perf record -e '{cycles,instructions}:S' and then browse context 43To show context switches in perf report sample context add --switch-events to perf record.