1PERF-BENCH(1)                     perf Manual                    PERF-BENCH(1)
2
3
4

NAME

6       perf-bench - General framework for benchmark suites
7

SYNOPSIS

9       perf bench [<common options>] <subsystem> <suite> [<options>]
10

DESCRIPTION

12       This perf bench command is a general framework for benchmark suites.
13

COMMON OPTIONS

15       -r, --repeat=
16           Specify amount of times to repeat the run (default 10).
17
18       -f, --format=
19           Specify format style. Current available format styles are:
20
21       default
22           Default style. This is mainly for human reading.
23
24
25           .ft C
26           % perf bench sched pipe                      # with no style specified
27           (executing 1000000 pipe operations between two tasks)
28                   Total time:5.855 sec
29                           5.855061 usecs/op
30                           170792 ops/sec
31           .ft
32
33
34       simple
35           This simple style is friendly for automated processing by scripts.
36
37
38           .ft C
39           % perf bench --format=simple sched pipe      # specified simple
40           5.988
41           .ft
42
43

SUBSYSTEM

45       sched
46           Scheduler and IPC mechanisms.
47
48       mem
49           Memory access performance.
50
51       numa
52           NUMA scheduling and MM benchmarks.
53
54       futex
55           Futex stressing benchmarks.
56
57       all
58           All benchmark subsystems.
59
60   SUITES FOR sched
61       messaging
62           Suite for evaluating performance of scheduler and IPC mechanisms.
63           Based on hackbench by Rusty Russell.
64
65       Options of messaging
66           -p, --pipe
67               Use pipe() instead of socketpair()
68
69           -t, --thread
70               Be multi thread instead of multi process
71
72           -g, --group=
73               Specify number of groups
74
75           -l, --loop=
76               Specify number of loops
77
78       Example of messaging
79               .ft C
80               % perf bench sched messaging                 # run with default
81               options (20 sender and receiver processes per group)
82               (10 groups == 400 processes run)
83
84                     Total time:0.308 sec
85
86               % perf bench sched messaging -t -g 20        # be multi-thread, with 20 groups
87               (20 sender and receiver threads per group)
88               (20 groups == 800 threads run)
89
90                     Total time:0.582 sec
91               .ft
92
93
94           pipe
95               Suite for pipe() system call. Based on pipe-test-1m.c by Ingo
96               Molnar.
97
98       Options of pipe
99           -l, --loop=
100               Specify number of loops.
101
102       Example of pipe
103               .ft C
104               % perf bench sched pipe
105               (executing 1000000 pipe operations between two tasks)
106
107                       Total time:8.091 sec
108                               8.091833 usecs/op
109                               123581 ops/sec
110
111               % perf bench sched pipe -l 1000              # loop 1000
112               (executing 1000 pipe operations between two tasks)
113
114                       Total time:0.016 sec
115                               16.948000 usecs/op
116                               59004 ops/sec
117               .ft
118
119
120   SUITES FOR mem
121       memcpy
122           Suite for evaluating performance of simple memory copy in various
123           ways.
124
125       Options of memcpy
126           -l, --length
127               Specify length of memory to copy (default: 1MB). Available
128               units are B, KB, MB, GB and TB (case insensitive).
129
130           -r, --routine
131               Specify routine to copy (default: default). Available routines
132               are depend on the architecture. On x86-64, x86-64-unrolled,
133               x86-64-movsq and x86-64-movsb are supported.
134
135           -i, --iterations
136               Repeat memcpy invocation this number of times.
137
138           -c, --clock
139               Use perf’s cpu-cycles event instead of gettimeofday syscall.
140
141           -o, --only-prefault
142               Show only the result with page faults before memcpy.
143
144           -n, --no-prefault
145               Show only the result without page faults before memcpy.
146
147           memset
148               Suite for evaluating performance of simple memory set in
149               various ways.
150
151       Options of memset
152           -l, --length
153               Specify length of memory to set (default: 1MB). Available units
154               are B, KB, MB, GB and TB (case insensitive).
155
156           -r, --routine
157               Specify routine to set (default: default). Available routines
158               are depend on the architecture. On x86-64, x86-64-unrolled,
159               x86-64-stosq and x86-64-stosb are supported.
160
161           -i, --iterations
162               Repeat memset invocation this number of times.
163
164           -c, --clock
165               Use perf’s cpu-cycles event instead of gettimeofday syscall.
166
167           -o, --only-prefault
168               Show only the result with page faults before memset.
169
170           -n, --no-prefault
171               Show only the result without page faults before memset.
172
173   SUITES FOR numa
174       mem
175           Suite for evaluating NUMA workloads.
176
177   SUITES FOR futex
178       hash
179           Suite for evaluating hash tables.
180
181       wake
182           Suite for evaluating wake calls.
183
184       requeue
185           Suite for evaluating requeue calls.
186

SEE ALSO

188       perf(1)
189
190
191
192perf                              06/18/2019                     PERF-BENCH(1)