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 5To not accumulate CPU time of children symbols add --no-children 6Customize output of perf script with: perf script -F event,ip,sym 7Generate a script for your data: perf script -g <lang> 8Save output of perf stat using: perf stat record <target workload> 9Create an archive with symtabs to analyse on other machine: perf archive 10Search options using a keyword: perf report -h <keyword> 11Use parent filter to see specific call path: perf report -p <regex> 12List events using substring match: perf list <keyword> 13To see list of saved events and attributes: perf evlist -v 14Use --symfs <dir> if your symbol files are in non-standard locations 15To see callchains in a more compact form: perf report -g folded 16To see call chains by final symbol taking CPU time (bottom up) use perf report -G 17Show individual samples with: perf script 18Limit to show entries above 5% only: perf report --percent-limit 5 19Profiling branch (mis)predictions with: perf record -b / perf report 20To show assembler sample context control flow use perf record -b / perf report --samples 10 and then browse context 21To adjust path to source files to local file system use perf report --prefix=... --prefix-strip=... 22Treat branches as callchains: perf record -b ... ; perf report --branch-history 23Show estimate cycles per function and IPC in annotate use perf record -b ... ; perf report --total-cycles 24To count events every 1000 msec: perf stat -I 1000 25Print event counts in machine readable CSV format with: perf stat -x\; 26If you have debuginfo enabled, try: perf report -s sym,srcline 27For memory address profiling, try: perf mem record / perf mem report 28For tracepoint events, try: perf report -s trace_fields 29To record callchains for each sample: perf record -g 30If call chains don't work try perf record --call-graph dwarf or --call-graph lbr 31To record every process run by a user: perf record -u <user> 32To show inline functions in call traces add --inline to perf report 33To not record events from perf itself add --exclude-perf 34Skip collecting build-id when recording: perf record -B 35To change sampling frequency to 100 Hz: perf record -F 100 36To show information about system the samples were collected on use perf report --header 37To only collect call graph on one event use perf record -e cpu/cpu-cycles,callgraph=1/,branches ; perf report --show-ref-call-graph 38To set sampling period of individual events use perf record -e cpu/cpu-cycles,period=100001/,cpu/branches,period=10001/ ... 39To group events which need to be collected together for accuracy use {}: perf record -e {cycles,branches}' ... 40To compute metrics for samples use perf record -e '{cycles,instructions}' ... ; perf script -F +metric 41See assembly instructions with percentage: perf annotate <symbol> 42If you prefer Intel style assembly, try: perf annotate -M intel 43When collecting LBR backtraces use --stitch-lbr to handle more than 32 deep entries: perf record --call-graph lbr ; perf report --stitch-lbr 44For hierarchical output, try: perf report --hierarchy 45Order by the overhead of source file name and line number: perf report -s srcline 46System-wide collection from all CPUs: perf record -a 47Show current config key-value pairs: perf config --list 48To collect Processor Trace with samples use perf record -e '{intel_pt//,cycles}' ; perf script --call-trace or --insn-trace --xed -F +ipc (remove --xed if no xed) 49To trace calls using Processor Trace use perf record -e intel_pt// ... ; perf script --call-trace. Then use perf script --time A-B --insn-trace to look at region of interest. 50To measure approximate function latency with Processor Trace use perf record -e intel_pt// ... ; perf script --call-ret-trace 51To trace only single function with Processor Trace use perf record --filter 'filter func @ program' -e intel_pt//u ./program ; perf script --insn-trace 52Show user configuration overrides: perf config --user --list 53To add Node.js USDT(User-Level Statically Defined Tracing): perf buildid-cache --add `which node` 54To analyze cache line scalability issues use perf c2c record ... ; perf c2c report 55To browse sample contexts use perf report --sample 10 and select in context menu 56To separate samples by time use perf report --sort time,overhead,sym 57To filter subset of samples with report or script add --time X-Y or --cpu A,B,C or --socket-filter ... 58To set sample time separation other than 100ms with --sort time use --time-quantum 59Add -I to perf record to sample register values, which will be visible in perf report sample context. 60To show IPC for sampling periods use perf record -e '{cycles,instructions}:S' and then browse context 61To show context switches in perf report sample context add --switch-events to perf record. 62To show time in nanoseconds in record/report add --ns 63To compare hot regions in two workloads use perf record -b -o file ... ; perf diff --stream file1 file2 64To compare scalability of two workload samples use perf diff -c ratio file1 file2 65