1perf-stat(1) 2============ 3 4NAME 5---- 6perf-stat - Run a command and gather performance counter statistics 7 8SYNOPSIS 9-------- 10[verse] 11'perf stat' [-e <EVENT> | --event=EVENT] [-a] <command> 12'perf stat' [-e <EVENT> | --event=EVENT] [-a] -- <command> [<options>] 13'perf stat' [-e <EVENT> | --event=EVENT] [-a] record [-o file] -- <command> [<options>] 14'perf stat' report [-i file] 15 16DESCRIPTION 17----------- 18This command runs a command and gathers performance counter statistics 19from it. 20 21 22OPTIONS 23------- 24<command>...:: 25 Any command you can specify in a shell. 26 27record:: 28 See STAT RECORD. 29 30report:: 31 See STAT REPORT. 32 33-e:: 34--event=:: 35 Select the PMU event. Selection can be: 36 37 - a symbolic event name (use 'perf list' to list all events) 38 39 - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a 40 hexadecimal event descriptor. 41 42 - a symbolically formed event like 'pmu/param1=0x3,param2/' where 43 param1 and param2 are defined as formats for the PMU in 44 /sys/bus/event_sources/devices/<pmu>/format/* 45 46 - a symbolically formed event like 'pmu/config=M,config1=N,config2=K/' 47 where M, N, K are numbers (in decimal, hex, octal format). 48 Acceptable values for each of 'config', 'config1' and 'config2' 49 parameters are defined by corresponding entries in 50 /sys/bus/event_sources/devices/<pmu>/format/* 51 52-i:: 53--no-inherit:: 54 child tasks do not inherit counters 55-p:: 56--pid=<pid>:: 57 stat events on existing process id (comma separated list) 58 59-t:: 60--tid=<tid>:: 61 stat events on existing thread id (comma separated list) 62 63 64-a:: 65--all-cpus:: 66 system-wide collection from all CPUs 67 68-c:: 69--scale:: 70 scale/normalize counter values 71 72-r:: 73--repeat=<n>:: 74 repeat command and print average + stddev (max: 100). 0 means forever. 75 76-B:: 77--big-num:: 78 print large numbers with thousands' separators according to locale 79 80-C:: 81--cpu=:: 82Count only on the list of CPUs provided. Multiple CPUs can be provided as a 83comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2. 84In per-thread mode, this option is ignored. The -a option is still necessary 85to activate system-wide monitoring. Default is to count on all CPUs. 86 87-A:: 88--no-aggr:: 89Do not aggregate counts across all monitored CPUs in system-wide mode (-a). 90This option is only valid in system-wide mode. 91 92-n:: 93--null:: 94 null run - don't start any counters 95 96-v:: 97--verbose:: 98 be more verbose (show counter open errors, etc) 99 100-x SEP:: 101--field-separator SEP:: 102print counts using a CSV-style output to make it easy to import directly into 103spreadsheets. Columns are separated by the string specified in SEP. 104 105-G name:: 106--cgroup name:: 107monitor only in the container (cgroup) called "name". This option is available only 108in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to 109container "name" are monitored when they run on the monitored CPUs. Multiple cgroups 110can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup 111to first event, second cgroup to second event and so on. It is possible to provide 112an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have 113corresponding events, i.e., they always refer to events defined earlier on the command 114line. 115 116-o file:: 117--output file:: 118Print the output into the designated file. 119 120--append:: 121Append to the output file designated with the -o option. Ignored if -o is not specified. 122 123--log-fd:: 124 125Log output to fd, instead of stderr. Complementary to --output, and mutually exclusive 126with it. --append may be used here. Examples: 127 3>results perf stat --log-fd 3 -- $cmd 128 3>>results perf stat --log-fd 3 --append -- $cmd 129 130--pre:: 131--post:: 132 Pre and post measurement hooks, e.g.: 133 134perf stat --repeat 10 --null --sync --pre 'make -s O=defconfig-build/clean' -- make -s -j64 O=defconfig-build/ bzImage 135 136-I msecs:: 137--interval-print msecs:: 138Print count deltas every N milliseconds (minimum: 10ms) 139The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals. Use with caution. 140 example: 'perf stat -I 1000 -e cycles -a sleep 5' 141 142--per-socket:: 143Aggregate counts per processor socket for system-wide mode measurements. This 144is a useful mode to detect imbalance between sockets. To enable this mode, 145use --per-socket in addition to -a. (system-wide). The output includes the 146socket number and the number of online processors on that socket. This is 147useful to gauge the amount of aggregation. 148 149--per-core:: 150Aggregate counts per physical processor for system-wide mode measurements. This 151is a useful mode to detect imbalance between physical cores. To enable this mode, 152use --per-core in addition to -a. (system-wide). The output includes the 153core number and the number of online logical processors on that physical processor. 154 155--per-thread:: 156Aggregate counts per monitored threads, when monitoring threads (-t option) 157or processes (-p option). 158 159-D msecs:: 160--delay msecs:: 161After starting the program, wait msecs before measuring. This is useful to 162filter out the startup phase of the program, which is often very different. 163 164-T:: 165--transaction:: 166 167Print statistics of transactional execution if supported. 168 169STAT RECORD 170----------- 171Stores stat data into perf data file. 172 173-o file:: 174--output file:: 175Output file name. 176 177STAT REPORT 178----------- 179Reads and reports stat data from perf data file. 180 181-i file:: 182--input file:: 183Input file name. 184 185--per-socket:: 186Aggregate counts per processor socket for system-wide mode measurements. 187 188--per-core:: 189Aggregate counts per physical processor for system-wide mode measurements. 190 191-A:: 192--no-aggr:: 193Do not aggregate counts across all monitored CPUs. 194 195 196EXAMPLES 197-------- 198 199$ perf stat -- make -j 200 201 Performance counter stats for 'make -j': 202 203 8117.370256 task clock ticks # 11.281 CPU utilization factor 204 678 context switches # 0.000 M/sec 205 133 CPU migrations # 0.000 M/sec 206 235724 pagefaults # 0.029 M/sec 207 24821162526 CPU cycles # 3057.784 M/sec 208 18687303457 instructions # 2302.138 M/sec 209 172158895 cache references # 21.209 M/sec 210 27075259 cache misses # 3.335 M/sec 211 212 Wall-clock time elapsed: 719.554352 msecs 213 214SEE ALSO 215-------- 216linkperf:perf-top[1], linkperf:perf-list[1] 217