261 lines · plain
1perf-bench(1)2=============3 4NAME5----6perf-bench - General framework for benchmark suites7 8SYNOPSIS9--------10[verse]11'perf bench' [<common options>] <subsystem> <suite> [<options>]12 13DESCRIPTION14-----------15This 'perf bench' command is a general framework for benchmark suites.16 17COMMON OPTIONS18--------------19-r::20--repeat=::21Specify number of times to repeat the run (default 10).22 23-f::24--format=::25Specify format style.26Current available format styles are:27 28'default'::29Default style. This is mainly for human reading.30---------------------31% perf bench sched pipe # with no style specified32(executing 1000000 pipe operations between two tasks)33 Total time:5.855 sec34 5.855061 usecs/op35 170792 ops/sec36---------------------37 38'simple'::39This simple style is friendly for automated40processing by scripts.41---------------------42% perf bench --format=simple sched pipe # specified simple435.98844---------------------45 46SUBSYSTEM47---------48 49'sched'::50 Scheduler and IPC mechanisms.51 52'syscall'::53 System call performance (throughput).54 55'mem'::56 Memory access performance.57 58'numa'::59 NUMA scheduling and MM benchmarks.60 61'futex'::62 Futex stressing benchmarks.63 64'epoll'::65 Eventpoll (epoll) stressing benchmarks.66 67'internals'::68 Benchmark internal perf functionality.69 70'uprobe'::71 Benchmark overhead of uprobe + BPF.72 73'all'::74 All benchmark subsystems.75 76SUITES FOR 'sched'77~~~~~~~~~~~~~~~~~~78*messaging*::79Suite for evaluating performance of scheduler and IPC mechanisms.80Based on hackbench by Rusty Russell.81 82Options of *messaging*83^^^^^^^^^^^^^^^^^^^^^^84-p::85--pipe::86Use pipe() instead of socketpair()87 88-t::89--thread::90Be multi thread instead of multi process91 92-g::93--group=::94Specify number of groups95 96-l::97--nr_loops=::98Specify number of loops99 100Example of *messaging*101^^^^^^^^^^^^^^^^^^^^^^102 103---------------------104% perf bench sched messaging # run with default105options (20 sender and receiver processes per group)106(10 groups == 400 processes run)107 108 Total time:0.308 sec109 110% perf bench sched messaging -t -g 20 # be multi-thread, with 20 groups111(20 sender and receiver threads per group)112(20 groups == 800 threads run)113 114 Total time:0.582 sec115---------------------116 117*pipe*::118Suite for pipe() system call.119Based on pipe-test-1m.c by Ingo Molnar.120 121Options of *pipe*122^^^^^^^^^^^^^^^^^123-l::124--loop=::125Specify number of loops.126 127-G::128--cgroups=::129Names of cgroups for sender and receiver, separated by a comma.130This is useful to check cgroup context switching overhead.131Note that perf doesn't create nor delete the cgroups, so users should132make sure that the cgroups exist and are accessible before use.133 134 135Example of *pipe*136^^^^^^^^^^^^^^^^^137 138---------------------139% perf bench sched pipe140(executing 1000000 pipe operations between two tasks)141 142 Total time:8.091 sec143 8.091833 usecs/op144 123581 ops/sec145 146% perf bench sched pipe -l 1000 # loop 1000147(executing 1000 pipe operations between two tasks)148 149 Total time:0.016 sec150 16.948000 usecs/op151 59004 ops/sec152 153% perf bench sched pipe -G AAA,BBB154(executing 1000000 pipe operations between cgroups)155# Running 'sched/pipe' benchmark:156# Executed 1000000 pipe operations between two processes157 158 Total time: 6.886 [sec]159 160 6.886208 usecs/op161 145217 ops/sec162 163---------------------164 165SUITES FOR 'syscall'166~~~~~~~~~~~~~~~~~~167*basic*::168Suite for evaluating performance of core system call throughput (both usecs/op and ops/sec metrics).169This uses a single thread simply doing getppid(2), which is a simple syscall where the result is not170cached by glibc.171 172 173SUITES FOR 'mem'174~~~~~~~~~~~~~~~~175*memcpy*::176Suite for evaluating performance of simple memory copy in various ways.177 178Options of *memcpy*179^^^^^^^^^^^^^^^^^^^180-l::181--size::182Specify size of memory to copy (default: 1MB).183Available units are B, KB, MB, GB and TB (case insensitive).184 185-f::186--function::187Specify function to copy (default: default).188Available functions are depend on the architecture.189On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.190 191-l::192--nr_loops::193Repeat memcpy invocation this number of times.194 195-c::196--cycles::197Use perf's cpu-cycles event instead of gettimeofday syscall.198 199*memset*::200Suite for evaluating performance of simple memory set in various ways.201 202Options of *memset*203^^^^^^^^^^^^^^^^^^^204-l::205--size::206Specify size of memory to set (default: 1MB).207Available units are B, KB, MB, GB and TB (case insensitive).208 209-f::210--function::211Specify function to set (default: default).212Available functions are depend on the architecture.213On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.214 215-l::216--nr_loops::217Repeat memset invocation this number of times.218 219-c::220--cycles::221Use perf's cpu-cycles event instead of gettimeofday syscall.222 223SUITES FOR 'numa'224~~~~~~~~~~~~~~~~~225*mem*::226Suite for evaluating NUMA workloads.227 228SUITES FOR 'futex'229~~~~~~~~~~~~~~~~~~230*hash*::231Suite for evaluating hash tables.232 233*wake*::234Suite for evaluating wake calls.235 236*wake-parallel*::237Suite for evaluating parallel wake calls.238 239*requeue*::240Suite for evaluating requeue calls.241 242*lock-pi*::243Suite for evaluating futex lock_pi calls.244 245SUITES FOR 'epoll'246~~~~~~~~~~~~~~~~~~247*wait*::248Suite for evaluating concurrent epoll_wait calls.249 250*ctl*::251Suite for evaluating multiple epoll_ctl calls.252 253SUITES FOR 'internals'254~~~~~~~~~~~~~~~~~~~~~~255*synthesize*::256Suite for evaluating perf's event synthesis performance.257 258SEE ALSO259--------260linkperf:perf[1]261