tools/perf/Documentation/perf-bench.txt

   1 perf-bench(1)
   2 =============
   3
   4 NAME
   5 ----
   6 perf-bench - General framework for benchmark suites
   7
   8 SYNOPSIS
   9 --------
  10 [verse]
  11 'perf bench' [<common options>] <subsystem> <suite> [<options>]
  12
  13 DESCRIPTION
  14 -----------
  15 This 'perf bench' command is a general framework for benchmark suites.
  16
  17 COMMON OPTIONS
  18 --------------
  19 -r::
  20 --repeat=::
  21 Specify amount of times to repeat the run (default 10).
  22
  23 -f::
  24 --format=::
  25 Specify format style.
  26 Current available format styles are:
  27
  28 'default'::
  29 Default style. This is mainly for human reading.
  30 ---------------------
  31 % perf bench sched pipe                      # with no style specified
  32 (executing 1000000 pipe operations between two tasks)
  33         Total time:5.855 sec
  34                 5.855061 usecs/op
  35                 170792 ops/sec
  36 ---------------------
  37
  38 'simple'::
  39 This simple style is friendly for automated
  40 processing by scripts.
  41 ---------------------
  42 % perf bench --format=simple sched pipe      # specified simple
  43 5.988
  44 ---------------------
  45
  46 SUBSYSTEM
  47 ---------
  48
  49 'sched'::
  50         Scheduler and IPC mechanisms.
  51
  52 'mem'::
  53         Memory access performance.
  54
  55 'numa'::
  56         NUMA scheduling and MM benchmarks.
  57
  58 'futex'::
  59         Futex stressing benchmarks.
  60
  61 'epoll'::
  62         Eventpoll (epoll) stressing benchmarks.
  63
  64 'internals'::
  65         Benchmark internal perf functionality.
  66
  67 'all'::
  68         All benchmark subsystems.
  69
  70 SUITES FOR 'sched'
  71 ~~~~~~~~~~~~~~~~~~
  72 *messaging*::
  73 Suite for evaluating performance of scheduler and IPC mechanisms.
  74 Based on hackbench by Rusty Russell.
  75
  76 Options of *messaging*
  77 ^^^^^^^^^^^^^^^^^^^^^^
  78 -p::
  79 --pipe::
  80 Use pipe() instead of socketpair()
  81
  82 -t::
  83 --thread::
  84 Be multi thread instead of multi process
  85
  86 -g::
  87 --group=::
  88 Specify number of groups
  89
  90 -l::
  91 --nr_loops=::
  92 Specify number of loops
  93
  94 Example of *messaging*
  95 ^^^^^^^^^^^^^^^^^^^^^^
  96
  97 ---------------------
  98 % perf bench sched messaging                 # run with default
  99 options (20 sender and receiver processes per group)
 100 (10 groups == 400 processes run)
 101
 102       Total time:0.308 sec
 103
 104 % perf bench sched messaging -t -g 20        # be multi-thread, with 20 groups
 105 (20 sender and receiver threads per group)
 106 (20 groups == 800 threads run)
 107
 108       Total time:0.582 sec
 109 ---------------------
 110
 111 *pipe*::
 112 Suite for pipe() system call.
 113 Based on pipe-test-1m.c by Ingo Molnar.
 114
 115 Options of *pipe*
 116 ^^^^^^^^^^^^^^^^^
 117 -l::
 118 --loop=::
 119 Specify number of loops.
 120
 121 Example of *pipe*
 122 ^^^^^^^^^^^^^^^^^
 123
 124 ---------------------
 125 % perf bench sched pipe
 126 (executing 1000000 pipe operations between two tasks)
 127
 128         Total time:8.091 sec
 129                 8.091833 usecs/op
 130                 123581 ops/sec
 131
 132 % perf bench sched pipe -l 1000              # loop 1000
 133 (executing 1000 pipe operations between two tasks)
 134
 135         Total time:0.016 sec
 136                 16.948000 usecs/op
 137                 59004 ops/sec
 138 ---------------------
 139
 140 SUITES FOR 'mem'
 141 ~~~~~~~~~~~~~~~~
 142 *memcpy*::
 143 Suite for evaluating performance of simple memory copy in various ways.
 144
 145 Options of *memcpy*
 146 ^^^^^^^^^^^^^^^^^^^
 147 -l::
 148 --size::
 149 Specify size of memory to copy (default: 1MB).
 150 Available units are B, KB, MB, GB and TB (case insensitive).
 151
 152 -f::
 153 --function::
 154 Specify function to copy (default: default).
 155 Available functions are depend on the architecture.
 156 On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.
 157
 158 -l::
 159 --nr_loops::
 160 Repeat memcpy invocation this number of times.
 161
 162 -c::
 163 --cycles::
 164 Use perf's cpu-cycles event instead of gettimeofday syscall.
 165
 166 *memset*::
 167 Suite for evaluating performance of simple memory set in various ways.
 168
 169 Options of *memset*
 170 ^^^^^^^^^^^^^^^^^^^
 171 -l::
 172 --size::
 173 Specify size of memory to set (default: 1MB).
 174 Available units are B, KB, MB, GB and TB (case insensitive).
 175
 176 -f::
 177 --function::
 178 Specify function to set (default: default).
 179 Available functions are depend on the architecture.
 180 On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.
 181
 182 -l::
 183 --nr_loops::
 184 Repeat memset invocation this number of times.
 185
 186 -c::
 187 --cycles::
 188 Use perf's cpu-cycles event instead of gettimeofday syscall.
 189
 190 SUITES FOR 'numa'
 191 ~~~~~~~~~~~~~~~~~
 192 *mem*::
 193 Suite for evaluating NUMA workloads.
 194
 195 SUITES FOR 'futex'
 196 ~~~~~~~~~~~~~~~~~~
 197 *hash*::
 198 Suite for evaluating hash tables.
 199
 200 *wake*::
 201 Suite for evaluating wake calls.
 202
 203 *wake-parallel*::
 204 Suite for evaluating parallel wake calls.
 205
 206 *requeue*::
 207 Suite for evaluating requeue calls.
 208
 209 *lock-pi*::
 210 Suite for evaluating futex lock_pi calls.
 211
 212 SUITES FOR 'epoll'
 213 ~~~~~~~~~~~~~~~~~~
 214 *wait*::
 215 Suite for evaluating concurrent epoll_wait calls.
 216
 217 *ctl*::
 218 Suite for evaluating multiple epoll_ctl calls.
 219
 220 SUITES FOR 'internals'
 221 ~~~~~~~~~~~~~~~~~~~~~~
 222 *synthesize*::
 223 Suite for evaluating perf's event synthesis performance.
 224
 225 SEE ALSO
 226 --------
 227 linkperf:perf[1]