6 perf-bench - General framework for benchmark suites
11 'perf bench' [<common options>] <subsystem> <suite> [<options>]
15 This 'perf bench' command is a general framework for benchmark suites.
21 Specify amount of times to repeat the run (default 10).
26 Current available format styles are:
29 Default style. This is mainly for human reading.
31 % perf bench sched pipe # with no style specified
32 (executing 1000000 pipe operations between two tasks)
39 This simple style is friendly for automated
40 processing by scripts.
42 % perf bench --format=simple sched pipe # specified simple
50 Scheduler and IPC mechanisms.
53 Memory access performance.
56 NUMA scheduling and MM benchmarks.
59 Futex stressing benchmarks.
62 Eventpoll (epoll) stressing benchmarks.
65 All benchmark subsystems.
70 Suite for evaluating performance of scheduler and IPC mechanisms.
71 Based on hackbench by Rusty Russell.
73 Options of *messaging*
74 ^^^^^^^^^^^^^^^^^^^^^^
77 Use pipe() instead of socketpair()
81 Be multi thread instead of multi process
85 Specify number of groups
89 Specify number of loops
91 Example of *messaging*
92 ^^^^^^^^^^^^^^^^^^^^^^
95 % perf bench sched messaging # run with default
96 options (20 sender and receiver processes per group)
97 (10 groups == 400 processes run)
101 % perf bench sched messaging -t -g 20 # be multi-thread, with 20 groups
102 (20 sender and receiver threads per group)
103 (20 groups == 800 threads run)
106 ---------------------
109 Suite for pipe() system call.
110 Based on pipe-test-1m.c by Ingo Molnar.
116 Specify number of loops.
121 ---------------------
122 % perf bench sched pipe
123 (executing 1000000 pipe operations between two tasks)
129 % perf bench sched pipe -l 1000 # loop 1000
130 (executing 1000 pipe operations between two tasks)
135 ---------------------
140 Suite for evaluating performance of simple memory copy in various ways.
146 Specify size of memory to copy (default: 1MB).
147 Available units are B, KB, MB, GB and TB (case insensitive).
151 Specify function to copy (default: default).
152 Available functions are depend on the architecture.
153 On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.
157 Repeat memcpy invocation this number of times.
161 Use perf's cpu-cycles event instead of gettimeofday syscall.
164 Suite for evaluating performance of simple memory set in various ways.
170 Specify size of memory to set (default: 1MB).
171 Available units are B, KB, MB, GB and TB (case insensitive).
175 Specify function to set (default: default).
176 Available functions are depend on the architecture.
177 On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.
181 Repeat memset invocation this number of times.
185 Use perf's cpu-cycles event instead of gettimeofday syscall.
190 Suite for evaluating NUMA workloads.
195 Suite for evaluating hash tables.
198 Suite for evaluating wake calls.
201 Suite for evaluating parallel wake calls.
204 Suite for evaluating requeue calls.
207 Suite for evaluating futex lock_pi calls.
212 Suite for evaluating concurrent epoll_wait calls.
215 Suite for evaluating multiple epoll_ctl calls.