1 perf-bench(1) 1 perf-bench(1)
2 ============= 2 =============
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-bench - General framework for benchmark s 6 perf-bench - General framework for benchmark suites
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf bench' [<common options>] <subsystem> <s 11 'perf bench' [<common options>] <subsystem> <suite> [<options>]
12 12
13 DESCRIPTION 13 DESCRIPTION
14 ----------- 14 -----------
15 This 'perf bench' command is a general framewo 15 This 'perf bench' command is a general framework for benchmark suites.
16 16
17 COMMON OPTIONS 17 COMMON OPTIONS
18 -------------- 18 --------------
19 -r:: 19 -r::
20 --repeat=:: 20 --repeat=::
21 Specify number of times to repeat the run (def !! 21 Specify amount of times to repeat the run (default 10).
22 22
23 -f:: 23 -f::
24 --format=:: 24 --format=::
25 Specify format style. 25 Specify format style.
26 Current available format styles are: 26 Current available format styles are:
27 27
28 'default':: 28 'default'::
29 Default style. This is mainly for human readin 29 Default style. This is mainly for human reading.
30 --------------------- 30 ---------------------
31 % perf bench sched pipe # 31 % perf bench sched pipe # with no style specified
32 (executing 1000000 pipe operations between two 32 (executing 1000000 pipe operations between two tasks)
33 Total time:5.855 sec 33 Total time:5.855 sec
34 5.855061 usecs/op 34 5.855061 usecs/op
35 170792 ops/sec 35 170792 ops/sec
36 --------------------- 36 ---------------------
37 37
38 'simple':: 38 'simple'::
39 This simple style is friendly for automated 39 This simple style is friendly for automated
40 processing by scripts. 40 processing by scripts.
41 --------------------- 41 ---------------------
42 % perf bench --format=simple sched pipe # 42 % perf bench --format=simple sched pipe # specified simple
43 5.988 43 5.988
44 --------------------- 44 ---------------------
45 45
46 SUBSYSTEM 46 SUBSYSTEM
47 --------- 47 ---------
48 48
49 'sched':: 49 'sched'::
50 Scheduler and IPC mechanisms. 50 Scheduler and IPC mechanisms.
51 51
52 'syscall':: 52 'syscall'::
53 System call performance (throughput). 53 System call performance (throughput).
54 54
55 'mem':: 55 'mem'::
56 Memory access performance. 56 Memory access performance.
57 57
58 'numa':: 58 'numa'::
59 NUMA scheduling and MM benchmarks. 59 NUMA scheduling and MM benchmarks.
60 60
61 'futex':: 61 'futex'::
62 Futex stressing benchmarks. 62 Futex stressing benchmarks.
63 63
64 'epoll':: 64 'epoll'::
65 Eventpoll (epoll) stressing benchmarks 65 Eventpoll (epoll) stressing benchmarks.
66 66
67 'internals':: 67 'internals'::
68 Benchmark internal perf functionality. 68 Benchmark internal perf functionality.
69 69
70 'uprobe':: <<
71 Benchmark overhead of uprobe + BPF. <<
72 <<
73 'all':: 70 'all'::
74 All benchmark subsystems. 71 All benchmark subsystems.
75 72
76 SUITES FOR 'sched' 73 SUITES FOR 'sched'
77 ~~~~~~~~~~~~~~~~~~ 74 ~~~~~~~~~~~~~~~~~~
78 *messaging*:: 75 *messaging*::
79 Suite for evaluating performance of scheduler 76 Suite for evaluating performance of scheduler and IPC mechanisms.
80 Based on hackbench by Rusty Russell. 77 Based on hackbench by Rusty Russell.
81 78
82 Options of *messaging* 79 Options of *messaging*
83 ^^^^^^^^^^^^^^^^^^^^^^ 80 ^^^^^^^^^^^^^^^^^^^^^^
84 -p:: 81 -p::
85 --pipe:: 82 --pipe::
86 Use pipe() instead of socketpair() 83 Use pipe() instead of socketpair()
87 84
88 -t:: 85 -t::
89 --thread:: 86 --thread::
90 Be multi thread instead of multi process 87 Be multi thread instead of multi process
91 88
92 -g:: 89 -g::
93 --group=:: 90 --group=::
94 Specify number of groups 91 Specify number of groups
95 92
96 -l:: 93 -l::
97 --nr_loops=:: 94 --nr_loops=::
98 Specify number of loops 95 Specify number of loops
99 96
100 Example of *messaging* 97 Example of *messaging*
101 ^^^^^^^^^^^^^^^^^^^^^^ 98 ^^^^^^^^^^^^^^^^^^^^^^
102 99
103 --------------------- 100 ---------------------
104 % perf bench sched messaging # 101 % perf bench sched messaging # run with default
105 options (20 sender and receiver processes per 102 options (20 sender and receiver processes per group)
106 (10 groups == 400 processes run) 103 (10 groups == 400 processes run)
107 104
108 Total time:0.308 sec 105 Total time:0.308 sec
109 106
110 % perf bench sched messaging -t -g 20 # 107 % perf bench sched messaging -t -g 20 # be multi-thread, with 20 groups
111 (20 sender and receiver threads per group) 108 (20 sender and receiver threads per group)
112 (20 groups == 800 threads run) 109 (20 groups == 800 threads run)
113 110
114 Total time:0.582 sec 111 Total time:0.582 sec
115 --------------------- 112 ---------------------
116 113
117 *pipe*:: 114 *pipe*::
118 Suite for pipe() system call. 115 Suite for pipe() system call.
119 Based on pipe-test-1m.c by Ingo Molnar. 116 Based on pipe-test-1m.c by Ingo Molnar.
120 117
121 Options of *pipe* 118 Options of *pipe*
122 ^^^^^^^^^^^^^^^^^ 119 ^^^^^^^^^^^^^^^^^
123 -l:: 120 -l::
124 --loop=:: 121 --loop=::
125 Specify number of loops. 122 Specify number of loops.
126 123
127 -G:: <<
128 --cgroups=:: <<
129 Names of cgroups for sender and receiver, sepa <<
130 This is useful to check cgroup context switchi <<
131 Note that perf doesn't create nor delete the c <<
132 make sure that the cgroups exist and are acces <<
133 <<
134 <<
135 Example of *pipe* 124 Example of *pipe*
136 ^^^^^^^^^^^^^^^^^ 125 ^^^^^^^^^^^^^^^^^
137 126
138 --------------------- 127 ---------------------
139 % perf bench sched pipe 128 % perf bench sched pipe
140 (executing 1000000 pipe operations between two 129 (executing 1000000 pipe operations between two tasks)
141 130
142 Total time:8.091 sec 131 Total time:8.091 sec
143 8.091833 usecs/op 132 8.091833 usecs/op
144 123581 ops/sec 133 123581 ops/sec
145 134
146 % perf bench sched pipe -l 1000 # 135 % perf bench sched pipe -l 1000 # loop 1000
147 (executing 1000 pipe operations between two ta 136 (executing 1000 pipe operations between two tasks)
148 137
149 Total time:0.016 sec 138 Total time:0.016 sec
150 16.948000 usecs/op 139 16.948000 usecs/op
151 59004 ops/sec 140 59004 ops/sec
152 <<
153 % perf bench sched pipe -G AAA,BBB <<
154 (executing 1000000 pipe operations between cgr <<
155 # Running 'sched/pipe' benchmark: <<
156 # Executed 1000000 pipe operations between two <<
157 <<
158 Total time: 6.886 [sec] <<
159 <<
160 6.886208 usecs/op <<
161 145217 ops/sec <<
162 <<
163 --------------------- 141 ---------------------
164 142
165 SUITES FOR 'syscall' 143 SUITES FOR 'syscall'
166 ~~~~~~~~~~~~~~~~~~ 144 ~~~~~~~~~~~~~~~~~~
167 *basic*:: 145 *basic*::
168 Suite for evaluating performance of core syste 146 Suite for evaluating performance of core system call throughput (both usecs/op and ops/sec metrics).
169 This uses a single thread simply doing getppid 147 This uses a single thread simply doing getppid(2), which is a simple syscall where the result is not
170 cached by glibc. 148 cached by glibc.
171 149
172 150
173 SUITES FOR 'mem' 151 SUITES FOR 'mem'
174 ~~~~~~~~~~~~~~~~ 152 ~~~~~~~~~~~~~~~~
175 *memcpy*:: 153 *memcpy*::
176 Suite for evaluating performance of simple mem 154 Suite for evaluating performance of simple memory copy in various ways.
177 155
178 Options of *memcpy* 156 Options of *memcpy*
179 ^^^^^^^^^^^^^^^^^^^ 157 ^^^^^^^^^^^^^^^^^^^
180 -l:: 158 -l::
181 --size:: 159 --size::
182 Specify size of memory to copy (default: 1MB). 160 Specify size of memory to copy (default: 1MB).
183 Available units are B, KB, MB, GB and TB (case 161 Available units are B, KB, MB, GB and TB (case insensitive).
184 162
185 -f:: 163 -f::
186 --function:: 164 --function::
187 Specify function to copy (default: default). 165 Specify function to copy (default: default).
188 Available functions are depend on the architec 166 Available functions are depend on the architecture.
189 On x86-64, x86-64-unrolled, x86-64-movsq and x 167 On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.
190 168
191 -l:: 169 -l::
192 --nr_loops:: 170 --nr_loops::
193 Repeat memcpy invocation this number of times. 171 Repeat memcpy invocation this number of times.
194 172
195 -c:: 173 -c::
196 --cycles:: 174 --cycles::
197 Use perf's cpu-cycles event instead of gettime 175 Use perf's cpu-cycles event instead of gettimeofday syscall.
198 176
199 *memset*:: 177 *memset*::
200 Suite for evaluating performance of simple mem 178 Suite for evaluating performance of simple memory set in various ways.
201 179
202 Options of *memset* 180 Options of *memset*
203 ^^^^^^^^^^^^^^^^^^^ 181 ^^^^^^^^^^^^^^^^^^^
204 -l:: 182 -l::
205 --size:: 183 --size::
206 Specify size of memory to set (default: 1MB). 184 Specify size of memory to set (default: 1MB).
207 Available units are B, KB, MB, GB and TB (case 185 Available units are B, KB, MB, GB and TB (case insensitive).
208 186
209 -f:: 187 -f::
210 --function:: 188 --function::
211 Specify function to set (default: default). 189 Specify function to set (default: default).
212 Available functions are depend on the architec 190 Available functions are depend on the architecture.
213 On x86-64, x86-64-unrolled, x86-64-stosq and x 191 On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.
214 192
215 -l:: 193 -l::
216 --nr_loops:: 194 --nr_loops::
217 Repeat memset invocation this number of times. 195 Repeat memset invocation this number of times.
218 196
219 -c:: 197 -c::
220 --cycles:: 198 --cycles::
221 Use perf's cpu-cycles event instead of gettime 199 Use perf's cpu-cycles event instead of gettimeofday syscall.
222 200
223 SUITES FOR 'numa' 201 SUITES FOR 'numa'
224 ~~~~~~~~~~~~~~~~~ 202 ~~~~~~~~~~~~~~~~~
225 *mem*:: 203 *mem*::
226 Suite for evaluating NUMA workloads. 204 Suite for evaluating NUMA workloads.
227 205
228 SUITES FOR 'futex' 206 SUITES FOR 'futex'
229 ~~~~~~~~~~~~~~~~~~ 207 ~~~~~~~~~~~~~~~~~~
230 *hash*:: 208 *hash*::
231 Suite for evaluating hash tables. 209 Suite for evaluating hash tables.
232 210
233 *wake*:: 211 *wake*::
234 Suite for evaluating wake calls. 212 Suite for evaluating wake calls.
235 213
236 *wake-parallel*:: 214 *wake-parallel*::
237 Suite for evaluating parallel wake calls. 215 Suite for evaluating parallel wake calls.
238 216
239 *requeue*:: 217 *requeue*::
240 Suite for evaluating requeue calls. 218 Suite for evaluating requeue calls.
241 219
242 *lock-pi*:: 220 *lock-pi*::
243 Suite for evaluating futex lock_pi calls. 221 Suite for evaluating futex lock_pi calls.
244 222
245 SUITES FOR 'epoll' 223 SUITES FOR 'epoll'
246 ~~~~~~~~~~~~~~~~~~ 224 ~~~~~~~~~~~~~~~~~~
247 *wait*:: 225 *wait*::
248 Suite for evaluating concurrent epoll_wait cal 226 Suite for evaluating concurrent epoll_wait calls.
249 227
250 *ctl*:: 228 *ctl*::
251 Suite for evaluating multiple epoll_ctl calls. 229 Suite for evaluating multiple epoll_ctl calls.
252 230
253 SUITES FOR 'internals' 231 SUITES FOR 'internals'
254 ~~~~~~~~~~~~~~~~~~~~~~ 232 ~~~~~~~~~~~~~~~~~~~~~~
255 *synthesize*:: 233 *synthesize*::
256 Suite for evaluating perf's event synthesis pe 234 Suite for evaluating perf's event synthesis performance.
257 235
258 SEE ALSO 236 SEE ALSO
259 -------- 237 --------
260 linkperf:perf[1] 238 linkperf:perf[1]
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.