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 All benchmark subsystems.
67 Suite for evaluating performance of scheduler and IPC mechanisms.
68 Based on hackbench by Rusty Russell.
70 Options of *messaging*
71 ^^^^^^^^^^^^^^^^^^^^^^
74 Use pipe() instead of socketpair()
78 Be multi thread instead of multi process
82 Specify number of groups
86 Specify number of loops
88 Example of *messaging*
89 ^^^^^^^^^^^^^^^^^^^^^^
92 % perf bench sched messaging # run with default
93 options (20 sender and receiver processes per group)
94 (10 groups == 400 processes run)
98 % perf bench sched messaging -t -g 20 # be multi-thread, with 20 groups
99 (20 sender and receiver threads per group)
100 (20 groups == 800 threads run)
103 ---------------------
106 Suite for pipe() system call.
107 Based on pipe-test-1m.c by Ingo Molnar.
113 Specify number of loops.
118 ---------------------
119 % perf bench sched pipe
120 (executing 1000000 pipe operations between two tasks)
126 % perf bench sched pipe -l 1000 # loop 1000
127 (executing 1000 pipe operations between two tasks)
132 ---------------------
137 Suite for evaluating performance of simple memory copy in various ways.
143 Specify length of memory to copy (default: 1MB).
144 Available units are B, KB, MB, GB and TB (case insensitive).
148 Specify routine to copy (default: default).
149 Available routines are depend on the architecture.
150 On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.
154 Repeat memcpy invocation this number of times.
158 Use perf's cpu-cycles event instead of gettimeofday syscall.
162 Show only the result with page faults before memcpy.
166 Show only the result without page faults before memcpy.
169 Suite for evaluating performance of simple memory set in various ways.
175 Specify length of memory to set (default: 1MB).
176 Available units are B, KB, MB, GB and TB (case insensitive).
180 Specify routine to set (default: default).
181 Available routines are depend on the architecture.
182 On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.
186 Repeat memset invocation this number of times.
190 Use perf's cpu-cycles event instead of gettimeofday syscall.
194 Show only the result with page faults before memset.
198 Show only the result without page faults before memset.
203 Suite for evaluating NUMA workloads.
208 Suite for evaluating hash tables.
211 Suite for evaluating wake calls.
214 Suite for evaluating requeue calls.