1 perf-stat(1) 1 perf-stat(1)
2 ============ 2 ============
3 3
4 NAME 4 NAME
5 ---- 5 ----
6 perf-stat - Run a command and gather performan 6 perf-stat - Run a command and gather performance counter statistics
7 7
8 SYNOPSIS 8 SYNOPSIS
9 -------- 9 --------
10 [verse] 10 [verse]
11 'perf stat' [-e <EVENT> | --event=EVENT] [-a] 11 'perf stat' [-e <EVENT> | --event=EVENT] [-a] <command>
12 'perf stat' [-e <EVENT> | --event=EVENT] [-a] !! 12 'perf stat' [-e <EVENT> | --event=EVENT] [-a] -- <command> [<options>]
13 'perf stat' [-e <EVENT> | --event=EVENT] [-a] !! 13 'perf stat' [-e <EVENT> | --event=EVENT] [-a] record [-o file] -- <command> [<options>]
14 'perf stat' report [-i file] 14 'perf stat' report [-i file]
15 15
16 DESCRIPTION 16 DESCRIPTION
17 ----------- 17 -----------
18 This command runs a command and gathers perfor 18 This command runs a command and gathers performance counter statistics
19 from it. 19 from it.
20 20
21 21
22 OPTIONS 22 OPTIONS
23 ------- 23 -------
24 <command>...:: 24 <command>...::
25 Any command you can specify in a shell 25 Any command you can specify in a shell.
26 26
27 record:: 27 record::
28 See STAT RECORD. 28 See STAT RECORD.
29 29
30 report:: 30 report::
31 See STAT REPORT. 31 See STAT REPORT.
32 32
33 -e:: 33 -e::
34 --event=:: 34 --event=::
35 Select the PMU event. Selection can be 35 Select the PMU event. Selection can be:
36 36
37 - a symbolic event name (use 'perf lis 37 - a symbolic event name (use 'perf list' to list all events)
38 38
39 - a raw PMU event in the form of rN wh !! 39 - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
40 that represents the raw register enc !! 40 hexadecimal event descriptor.
41 event control registers as described <<
42 /sys/bus/event_source/devices/cpu/fo <<
43 41
44 - a symbolic or raw PMU event followed 42 - a symbolic or raw PMU event followed by an optional colon
45 and a list of event modifiers, e.g., 43 and a list of event modifiers, e.g., cpu-cycles:p. See the
46 linkperf:perf-list[1] man page for d 44 linkperf:perf-list[1] man page for details on event modifiers.
47 45
48 - a symbolically formed event like 'pm 46 - a symbolically formed event like 'pmu/param1=0x3,param2/' where
49 param1 and param2 are defined as for 47 param1 and param2 are defined as formats for the PMU in
50 /sys/bus/event_source/devices/<pmu>/ 48 /sys/bus/event_source/devices/<pmu>/format/*
51 49
52 'percore' is a event qualifier that 50 'percore' is a event qualifier that sums up the event counts for both
53 hardware threads in a core. For exam 51 hardware threads in a core. For example:
54 perf stat -A -a -e cpu/event,percore 52 perf stat -A -a -e cpu/event,percore=1/,otherevent ...
55 53
56 - a symbolically formed event like 'pm 54 - a symbolically formed event like 'pmu/config=M,config1=N,config2=K/'
57 where M, N, K are numbers (in decima 55 where M, N, K are numbers (in decimal, hex, octal format).
58 Acceptable values for each of 'confi 56 Acceptable values for each of 'config', 'config1' and 'config2'
59 parameters are defined by correspond 57 parameters are defined by corresponding entries in
60 /sys/bus/event_source/devices/<pmu>/ 58 /sys/bus/event_source/devices/<pmu>/format/*
61 59
62 Note that the last two syntaxes suppor 60 Note that the last two syntaxes support prefix and glob matching in
63 the PMU name to simplify creation of e 61 the PMU name to simplify creation of events across multiple instances
64 of the same type of PMU in large syste 62 of the same type of PMU in large systems (e.g. memory controller PMUs).
65 Multiple PMU instances are typical for 63 Multiple PMU instances are typical for uncore PMUs, so the prefix
66 'uncore_' is also ignored when perform 64 'uncore_' is also ignored when performing this match.
67 65
68 66
69 -i:: 67 -i::
70 --no-inherit:: 68 --no-inherit::
71 child tasks do not inherit counters 69 child tasks do not inherit counters
72 -p:: 70 -p::
73 --pid=<pid>:: 71 --pid=<pid>::
74 stat events on existing process id (co 72 stat events on existing process id (comma separated list)
75 73
76 -t:: 74 -t::
77 --tid=<tid>:: 75 --tid=<tid>::
78 stat events on existing thread id (com 76 stat events on existing thread id (comma separated list)
79 77
80 -b:: 78 -b::
81 --bpf-prog:: 79 --bpf-prog::
82 stat events on existing bpf program id 80 stat events on existing bpf program id (comma separated list),
83 requiring root rights. bpftool-prog co 81 requiring root rights. bpftool-prog could be used to find program
84 id all bpf programs in the system. For 82 id all bpf programs in the system. For example:
85 83
86 # bpftool prog | head -n 1 84 # bpftool prog | head -n 1
87 17247: tracepoint name sys_enter tag 192d5 85 17247: tracepoint name sys_enter tag 192d548b9d754067 gpl
88 86
89 # perf stat -e cycles,instructions --bpf-pro 87 # perf stat -e cycles,instructions --bpf-prog 17247 --timeout 1000
90 88
91 Performance counter stats for 'BPF program( 89 Performance counter stats for 'BPF program(s) 17247':
92 90
93 85,967 cycles 91 85,967 cycles
94 28,982 instructions 92 28,982 instructions # 0.34 insn per cycle
95 93
96 1.102235068 seconds time elapsed 94 1.102235068 seconds time elapsed
97 95
98 --bpf-counters:: <<
99 Use BPF programs to aggregate readings <<
100 allows multiple perf-stat sessions tha <<
101 instructions, etc.) to share hardware <<
102 To use BPF programs on common events b <<
103 "perf config stat.bpf-counter-events=< <<
104 <<
105 --bpf-attr-map:: <<
106 With option "--bpf-counters", differen <<
107 information about shared BPF programs <<
108 Use "--bpf-attr-map" to specify the pa <<
109 The default path is /sys/fs/bpf/perf_a <<
110 <<
111 ifdef::HAVE_LIBPFM[] 96 ifdef::HAVE_LIBPFM[]
112 --pfm-events events:: 97 --pfm-events events::
113 Select a PMU event using libpfm4 syntax (see h 98 Select a PMU event using libpfm4 syntax (see http://perfmon2.sf.net)
114 including support for event filters. For examp 99 including support for event filters. For example '--pfm-events
115 inst_retired:any_p:u:c=1:i'. More than one eve 100 inst_retired:any_p:u:c=1:i'. More than one event can be passed to the
116 option using the comma separator. Hardware eve 101 option using the comma separator. Hardware events and generic hardware
117 events cannot be mixed together. The latter mu 102 events cannot be mixed together. The latter must be used with the -e
118 option. The -e option and this one can be mixe 103 option. The -e option and this one can be mixed and matched. Events
119 can be grouped using the {} notation. 104 can be grouped using the {} notation.
120 endif::HAVE_LIBPFM[] 105 endif::HAVE_LIBPFM[]
121 106
122 -a:: 107 -a::
123 --all-cpus:: 108 --all-cpus::
124 system-wide collection from all CPUs ( 109 system-wide collection from all CPUs (default if no target is specified)
125 110
126 --no-scale:: 111 --no-scale::
127 Don't scale/normalize counter values 112 Don't scale/normalize counter values
128 113
129 -d:: 114 -d::
130 --detailed:: 115 --detailed::
131 print more detailed statistics, can be 116 print more detailed statistics, can be specified up to 3 times
132 117
133 -d: detailed events, L1 an 118 -d: detailed events, L1 and LLC data cache
134 -d -d: more detailed events, dTLB 119 -d -d: more detailed events, dTLB and iTLB events
135 -d -d -d: very detailed events, addin 120 -d -d -d: very detailed events, adding prefetch events
136 121
137 -r:: 122 -r::
138 --repeat=<n>:: 123 --repeat=<n>::
139 repeat command and print average + std 124 repeat command and print average + stddev (max: 100). 0 means forever.
140 125
141 -B:: 126 -B::
142 --big-num:: 127 --big-num::
143 print large numbers with thousands' se 128 print large numbers with thousands' separators according to locale.
144 Enabled by default. Use "--no-big-num" 129 Enabled by default. Use "--no-big-num" to disable.
145 Default setting can be changed with "p 130 Default setting can be changed with "perf config stat.big-num=false".
146 131
147 -C:: 132 -C::
148 --cpu=:: 133 --cpu=::
149 Count only on the list of CPUs provided. Multi 134 Count only on the list of CPUs provided. Multiple CPUs can be provided as a
150 comma-separated list with no space: 0,1. Range 135 comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.
151 In per-thread mode, this option is ignored. Th 136 In per-thread mode, this option is ignored. The -a option is still necessary
152 to activate system-wide monitoring. Default is 137 to activate system-wide monitoring. Default is to count on all CPUs.
153 138
154 -A:: 139 -A::
155 --no-aggr:: 140 --no-aggr::
156 Do not aggregate counts across all monitored C 141 Do not aggregate counts across all monitored CPUs.
157 142
158 -n:: 143 -n::
159 --null:: 144 --null::
160 null run - Don't start any counters. !! 145 null run - don't start any counters
161 <<
162 This can be useful to measure just elapsed wal <<
163 raw overhead of perf stat itself, without runn <<
164 146
165 -v:: 147 -v::
166 --verbose:: 148 --verbose::
167 be more verbose (show counter open err 149 be more verbose (show counter open errors, etc)
168 150
169 -x SEP:: 151 -x SEP::
170 --field-separator SEP:: 152 --field-separator SEP::
171 print counts using a CSV-style output to make 153 print counts using a CSV-style output to make it easy to import directly into
172 spreadsheets. Columns are separated by the str 154 spreadsheets. Columns are separated by the string specified in SEP.
173 155
174 --table:: Display time for each run (-r option 156 --table:: Display time for each run (-r option), in a table format, e.g.:
175 157
176 $ perf stat --null -r 5 --table perf bench s 158 $ perf stat --null -r 5 --table perf bench sched pipe
177 159
178 Performance counter stats for 'perf bench s 160 Performance counter stats for 'perf bench sched pipe' (5 runs):
179 161
180 # Table of individual measurement 162 # Table of individual measurements:
181 5.189 (-0.293) # 163 5.189 (-0.293) #
182 5.189 (-0.294) # 164 5.189 (-0.294) #
183 5.186 (-0.296) # 165 5.186 (-0.296) #
184 5.663 (+0.181) ## 166 5.663 (+0.181) ##
185 6.186 (+0.703) #### 167 6.186 (+0.703) ####
186 168
187 # Final result: 169 # Final result:
188 5.483 +- 0.198 seconds time elaps 170 5.483 +- 0.198 seconds time elapsed ( +- 3.62% )
189 171
190 -G name:: 172 -G name::
191 --cgroup name:: 173 --cgroup name::
192 monitor only in the container (cgroup) called 174 monitor only in the container (cgroup) called "name". This option is available only
193 in per-cpu mode. The cgroup filesystem must be 175 in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to
194 container "name" are monitored when they run o 176 container "name" are monitored when they run on the monitored CPUs. Multiple cgroups
195 can be provided. Each cgroup is applied to the 177 can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup
196 to first event, second cgroup to second event 178 to first event, second cgroup to second event and so on. It is possible to provide
197 an empty cgroup (monitor all the time) using, 179 an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have
198 corresponding events, i.e., they always refer 180 corresponding events, i.e., they always refer to events defined earlier on the command
199 line. If the user wants to track multiple even 181 line. If the user wants to track multiple events for a specific cgroup, the user can
200 use '-e e1 -e e2 -G foo,foo' or just use '-e e 182 use '-e e1 -e e2 -G foo,foo' or just use '-e e1 -e e2 -G foo'.
201 183
202 If wanting to monitor, say, 'cycles' for a cgr 184 If wanting to monitor, say, 'cycles' for a cgroup and also for system wide, this
203 command line can be used: 'perf stat -e cycles 185 command line can be used: 'perf stat -e cycles -G cgroup_name -a -e cycles'.
204 186
205 --for-each-cgroup name:: 187 --for-each-cgroup name::
206 Expand event list for each cgroup in "name" (a 188 Expand event list for each cgroup in "name" (allow multiple cgroups separated
207 by comma). It also support regex patterns to 189 by comma). It also support regex patterns to match multiple groups. This has same
208 effect that repeating -e option and -G option 190 effect that repeating -e option and -G option for each event x name. This option
209 cannot be used with -G/--cgroup option. 191 cannot be used with -G/--cgroup option.
210 192
211 -o file:: 193 -o file::
212 --output file:: 194 --output file::
213 Print the output into the designated file. 195 Print the output into the designated file.
214 196
215 --append:: 197 --append::
216 Append to the output file designated with the 198 Append to the output file designated with the -o option. Ignored if -o is not specified.
217 199
218 --log-fd:: 200 --log-fd::
219 201
220 Log output to fd, instead of stderr. Compleme 202 Log output to fd, instead of stderr. Complementary to --output, and mutually exclusive
221 with it. --append may be used here. Examples 203 with it. --append may be used here. Examples:
222 3>results perf stat --log-fd 3 !! 204 3>results perf stat --log-fd 3 -- $cmd
223 3>>results perf stat --log-fd 3 --append !! 205 3>>results perf stat --log-fd 3 --append -- $cmd
224 206
225 --control=fifo:ctl-fifo[,ack-fifo]:: 207 --control=fifo:ctl-fifo[,ack-fifo]::
226 --control=fd:ctl-fd[,ack-fd]:: 208 --control=fd:ctl-fd[,ack-fd]::
227 ctl-fifo / ack-fifo are opened and used as ctl 209 ctl-fifo / ack-fifo are opened and used as ctl-fd / ack-fd as follows.
228 Listen on ctl-fd descriptor for command to con 210 Listen on ctl-fd descriptor for command to control measurement ('enable': enable events,
229 'disable': disable events). Measurements can b 211 'disable': disable events). Measurements can be started with events disabled using
230 --delay=-1 option. Optionally send control com 212 --delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor
231 to synchronize with the controlling process. E 213 to synchronize with the controlling process. Example of bash shell script to enable and
232 disable events during measurements: 214 disable events during measurements:
233 215
234 #!/bin/bash 216 #!/bin/bash
235 217
236 ctl_dir=/tmp/ 218 ctl_dir=/tmp/
237 219
238 ctl_fifo=${ctl_dir}perf_ctl.fifo 220 ctl_fifo=${ctl_dir}perf_ctl.fifo
239 test -p ${ctl_fifo} && unlink ${ctl_fifo} 221 test -p ${ctl_fifo} && unlink ${ctl_fifo}
240 mkfifo ${ctl_fifo} 222 mkfifo ${ctl_fifo}
241 exec {ctl_fd}<>${ctl_fifo} 223 exec {ctl_fd}<>${ctl_fifo}
242 224
243 ctl_ack_fifo=${ctl_dir}perf_ctl_ack.fifo 225 ctl_ack_fifo=${ctl_dir}perf_ctl_ack.fifo
244 test -p ${ctl_ack_fifo} && unlink ${ctl_ack_f 226 test -p ${ctl_ack_fifo} && unlink ${ctl_ack_fifo}
245 mkfifo ${ctl_ack_fifo} 227 mkfifo ${ctl_ack_fifo}
246 exec {ctl_fd_ack}<>${ctl_ack_fifo} 228 exec {ctl_fd_ack}<>${ctl_ack_fifo}
247 229
248 perf stat -D -1 -e cpu-cycles -a -I 1000 230 perf stat -D -1 -e cpu-cycles -a -I 1000 \
249 --control fd:${ctl_fd},${ctl_fd_ack 231 --control fd:${ctl_fd},${ctl_fd_ack} \
250 \-- sleep 30 & !! 232 -- sleep 30 &
251 perf_pid=$! 233 perf_pid=$!
252 234
253 sleep 5 && echo 'enable' >&${ctl_fd} && read 235 sleep 5 && echo 'enable' >&${ctl_fd} && read -u ${ctl_fd_ack} e1 && echo "enabled(${e1})"
254 sleep 10 && echo 'disable' >&${ctl_fd} && rea 236 sleep 10 && echo 'disable' >&${ctl_fd} && read -u ${ctl_fd_ack} d1 && echo "disabled(${d1})"
255 237
256 exec {ctl_fd_ack}>&- 238 exec {ctl_fd_ack}>&-
257 unlink ${ctl_ack_fifo} 239 unlink ${ctl_ack_fifo}
258 240
259 exec {ctl_fd}>&- 241 exec {ctl_fd}>&-
260 unlink ${ctl_fifo} 242 unlink ${ctl_fifo}
261 243
262 wait -n ${perf_pid} 244 wait -n ${perf_pid}
263 exit $? 245 exit $?
264 246
265 247
266 --pre:: 248 --pre::
267 --post:: 249 --post::
268 Pre and post measurement hooks, e.g.: 250 Pre and post measurement hooks, e.g.:
269 251
270 perf stat --repeat 10 --null --sync --pre 'mak !! 252 perf stat --repeat 10 --null --sync --pre 'make -s O=defconfig-build/clean' -- make -s -j64 O=defconfig-build/ bzImage
271 253
272 -I msecs:: 254 -I msecs::
273 --interval-print msecs:: 255 --interval-print msecs::
274 Print count deltas every N milliseconds (minim 256 Print count deltas every N milliseconds (minimum: 1ms)
275 The overhead percentage could be high in some 257 The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals. Use with caution.
276 example: 'perf stat -I 1000 -e cycles 258 example: 'perf stat -I 1000 -e cycles -a sleep 5'
277 259
278 If the metric exists, it is calculated by the 260 If the metric exists, it is calculated by the counts generated in this interval and the metric is printed after #.
279 261
280 --interval-count times:: 262 --interval-count times::
281 Print count deltas for fixed number of times. 263 Print count deltas for fixed number of times.
282 This option should be used together with "-I" 264 This option should be used together with "-I" option.
283 example: 'perf stat -I 1000 --interval 265 example: 'perf stat -I 1000 --interval-count 2 -e cycles -a'
284 266
285 --interval-clear:: 267 --interval-clear::
286 Clear the screen before next interval. 268 Clear the screen before next interval.
287 269
288 --timeout msecs:: 270 --timeout msecs::
289 Stop the 'perf stat' session and print count d 271 Stop the 'perf stat' session and print count deltas after N milliseconds (minimum: 10 ms).
290 This option is not supported with the "-I" opt 272 This option is not supported with the "-I" option.
291 example: 'perf stat --time 2000 -e cyc 273 example: 'perf stat --time 2000 -e cycles -a'
292 274
293 --metric-only:: 275 --metric-only::
294 Only print computed metrics. Print them in a s 276 Only print computed metrics. Print them in a single line.
295 Don't show any raw values. Not supported with 277 Don't show any raw values. Not supported with --per-thread.
296 278
297 --per-socket:: 279 --per-socket::
298 Aggregate counts per processor socket for syst 280 Aggregate counts per processor socket for system-wide mode measurements. This
299 is a useful mode to detect imbalance between s 281 is a useful mode to detect imbalance between sockets. To enable this mode,
300 use --per-socket in addition to -a. (system-wi 282 use --per-socket in addition to -a. (system-wide). The output includes the
301 socket number and the number of online process 283 socket number and the number of online processors on that socket. This is
302 useful to gauge the amount of aggregation. 284 useful to gauge the amount of aggregation.
303 285
304 --per-die:: 286 --per-die::
305 Aggregate counts per processor die for system- 287 Aggregate counts per processor die for system-wide mode measurements. This
306 is a useful mode to detect imbalance between d 288 is a useful mode to detect imbalance between dies. To enable this mode,
307 use --per-die in addition to -a. (system-wide) 289 use --per-die in addition to -a. (system-wide). The output includes the
308 die number and the number of online processors 290 die number and the number of online processors on that die. This is
309 useful to gauge the amount of aggregation. 291 useful to gauge the amount of aggregation.
310 292
311 --per-cluster:: <<
312 Aggregate counts per processor cluster for sys <<
313 is a useful mode to detect imbalance between c <<
314 use --per-cluster in addition to -a. (system-w <<
315 cluster number and the number of online proces <<
316 useful to gauge the amount of aggregation. The <<
317 related CPUs can be gotten from /sys/devices/s <<
318 <<
319 --per-cache:: <<
320 Aggregate counts per cache instance for system <<
321 default, the aggregation happens for the cache <<
322 in the system. To specify a particular level, <<
323 alongside the option in the format [Ll][1-9][0 <<
324 Using option "--per-cache=l3" or "--per-cache= <<
325 information at the boundary of the level 3 cac <<
326 <<
327 --per-core:: 293 --per-core::
328 Aggregate counts per physical processor for sy 294 Aggregate counts per physical processor for system-wide mode measurements. This
329 is a useful mode to detect imbalance between p 295 is a useful mode to detect imbalance between physical cores. To enable this mode,
330 use --per-core in addition to -a. (system-wide 296 use --per-core in addition to -a. (system-wide). The output includes the
331 core number and the number of online logical p 297 core number and the number of online logical processors on that physical processor.
332 298
333 --per-thread:: 299 --per-thread::
334 Aggregate counts per monitored threads, when m 300 Aggregate counts per monitored threads, when monitoring threads (-t option)
335 or processes (-p option). 301 or processes (-p option).
336 302
337 --per-node:: 303 --per-node::
338 Aggregate counts per NUMA nodes for system-wid 304 Aggregate counts per NUMA nodes for system-wide mode measurements. This
339 is a useful mode to detect imbalance between N 305 is a useful mode to detect imbalance between NUMA nodes. To enable this
340 mode, use --per-node in addition to -a. (syste 306 mode, use --per-node in addition to -a. (system-wide).
341 307
342 -D msecs:: 308 -D msecs::
343 --delay msecs:: 309 --delay msecs::
344 After starting the program, wait msecs before 310 After starting the program, wait msecs before measuring (-1: start with events
345 disabled). This is useful to filter out the st 311 disabled). This is useful to filter out the startup phase of the program,
346 which is often very different. 312 which is often very different.
347 313
348 -T:: 314 -T::
349 --transaction:: 315 --transaction::
350 316
351 Print statistics of transactional execution if 317 Print statistics of transactional execution if supported.
352 318
353 --metric-no-group:: 319 --metric-no-group::
354 By default, events to compute a metric are pla 320 By default, events to compute a metric are placed in weak groups. The
355 group tries to enforce scheduling all or none 321 group tries to enforce scheduling all or none of the events. The
356 --metric-no-group option places events outside 322 --metric-no-group option places events outside of groups and may
357 increase the chance of the event being schedul 323 increase the chance of the event being scheduled - leading to more
358 accuracy. However, as events may not be schedu 324 accuracy. However, as events may not be scheduled together accuracy
359 for metrics like instructions per cycle can be 325 for metrics like instructions per cycle can be lower - as both metrics
360 may no longer be being measured at the same ti 326 may no longer be being measured at the same time.
361 327
362 --metric-no-merge:: 328 --metric-no-merge::
363 By default metric events in different weak gro 329 By default metric events in different weak groups can be shared if one
364 group contains all the events needed by anothe 330 group contains all the events needed by another. In such cases one
365 group will be eliminated reducing event multip 331 group will be eliminated reducing event multiplexing and making it so
366 that certain groups of metrics sum to 100%. A 332 that certain groups of metrics sum to 100%. A downside to sharing a
367 group is that the group may require multiplexi 333 group is that the group may require multiplexing and so accuracy for a
368 small group that need not have multiplexing is 334 small group that need not have multiplexing is lowered. This option
369 forbids the event merging logic from sharing e 335 forbids the event merging logic from sharing events between groups and
370 may be used to increase accuracy in this case. 336 may be used to increase accuracy in this case.
371 337
372 --metric-no-threshold:: <<
373 Metric thresholds may increase the number of e <<
374 compute whether a metric has exceeded its thre <<
375 may not be desirable, for example, as the even <<
376 multiplexing. This option disables the adding <<
377 events for a metric. However, if there are suf <<
378 compute the threshold then the threshold is st <<
379 color the metric's computed value. <<
380 <<
381 --quiet:: 338 --quiet::
382 Don't print output, warnings or messages. This !! 339 Don't print output. This is useful with perf stat record below to only
383 record below to only write data to the perf.da !! 340 write data to the perf.data file.
384 341
385 STAT RECORD 342 STAT RECORD
386 ----------- 343 -----------
387 Stores stat data into perf data file. 344 Stores stat data into perf data file.
388 345
389 -o file:: 346 -o file::
390 --output file:: 347 --output file::
391 Output file name. 348 Output file name.
392 349
393 STAT REPORT 350 STAT REPORT
394 ----------- 351 -----------
395 Reads and reports stat data from perf data fil 352 Reads and reports stat data from perf data file.
396 353
397 -i file:: 354 -i file::
398 --input file:: 355 --input file::
399 Input file name. 356 Input file name.
400 357
401 --per-socket:: 358 --per-socket::
402 Aggregate counts per processor socket for syst 359 Aggregate counts per processor socket for system-wide mode measurements.
403 360
404 --per-die:: 361 --per-die::
405 Aggregate counts per processor die for system- 362 Aggregate counts per processor die for system-wide mode measurements.
406 363
407 --per-cluster:: <<
408 Aggregate counts perf processor cluster for sy <<
409 <<
410 --per-cache:: <<
411 Aggregate counts per cache instance for system <<
412 default, the aggregation happens for the cache <<
413 in the system. To specify a particular level, <<
414 alongside the option in the format [Ll][1-9][0 <<
415 option "--per-cache=l3" or "--per-cache=L3" wi <<
416 information at the boundary of the level 3 cac <<
417 <<
418 --per-core:: 364 --per-core::
419 Aggregate counts per physical processor for sy 365 Aggregate counts per physical processor for system-wide mode measurements.
420 366
421 -M:: 367 -M::
422 --metrics:: 368 --metrics::
423 Print metrics or metricgroups specified in a c 369 Print metrics or metricgroups specified in a comma separated list.
424 For a group all metrics from the group are add 370 For a group all metrics from the group are added.
425 The events from the metrics are automatically 371 The events from the metrics are automatically measured.
426 See perf list output for the possible metrics !! 372 See perf list output for the possble metrics and metricgroups.
427 <<
428 When threshold information is availabl <<
429 color red is used to signify a metric <<
430 while green shows it hasn't. The defau <<
431 no threshold information was available <<
432 couldn't be computed. <<
433 373
434 -A:: 374 -A::
435 --no-aggr:: 375 --no-aggr::
436 --no-merge:: !! 376 Do not aggregate counts across all monitored CPUs.
437 Do not aggregate/merge counts across monitored <<
438 <<
439 When multiple events are created from a single <<
440 stat will, by default, aggregate the event cou <<
441 in a single row. This option disables that beh <<
442 individual events and counts. <<
443 <<
444 Multiple events are created from a single even <<
445 <<
446 1. PID monitoring isn't requested and the syst <<
447 CPU. For example, a system with 8 SMT threa <<
448 opened on each thread and aggregation is pe <<
449 <<
450 2. Prefix or glob wildcard matching is used fo <<
451 example, multiple memory controller PMUs ma <<
452 suffix of _0, _1, etc. By default the event <<
453 combined if the PMU is specified without th <<
454 uncore_imc rather than uncore_imc_0. <<
455 <<
456 3. Aliases, which are listed immediately after <<
457 by perf list, are used. <<
458 <<
459 --hybrid-merge:: <<
460 Merge core event counts from all core PMUs. In <<
461 systems by default each core PMU will report i <<
462 separately. This option forces core PMU counts <<
463 a behavior closer to having a single CPU type <<
464 377
465 --topdown:: 378 --topdown::
466 Print top-down metrics supported by the CPU. T !! 379 Print complete top-down metrics supported by the CPU. This allows to
467 bottle necks in the CPU pipeline for CPU bound !! 380 determine bottle necks in the CPU pipeline for CPU bound workloads,
468 the cycles consumed down into frontend bound, !! 381 by breaking the cycles consumed down into frontend bound, backend bound,
469 speculation and retiring. !! 382 bad speculation and retiring.
470 383
471 Frontend bound means that the CPU cannot fetch 384 Frontend bound means that the CPU cannot fetch and decode instructions fast
472 enough. Backend bound means that computation o 385 enough. Backend bound means that computation or memory access is the bottle
473 neck. Bad Speculation means that the CPU waste 386 neck. Bad Speculation means that the CPU wasted cycles due to branch
474 mispredictions and similar issues. Retiring me 387 mispredictions and similar issues. Retiring means that the CPU computed without
475 an apparently bottleneck. The bottleneck is on 388 an apparently bottleneck. The bottleneck is only the real bottleneck
476 if the workload is actually bound by the CPU a 389 if the workload is actually bound by the CPU and not by something else.
477 390
478 For best results it is usually a good idea to 391 For best results it is usually a good idea to use it with interval
479 mode like -I 1000, as the bottleneck of worklo 392 mode like -I 1000, as the bottleneck of workloads can change often.
480 393
481 This enables --metric-only, unless overridden 394 This enables --metric-only, unless overridden with --no-metric-only.
482 395
483 The following restrictions only apply to older 396 The following restrictions only apply to older Intel CPUs and Atom,
484 on newer CPUs (IceLake and later) TopDown can 397 on newer CPUs (IceLake and later) TopDown can be collected for any thread:
485 398
486 The top down metrics are collected per core in 399 The top down metrics are collected per core instead of per
487 CPU thread. Per core mode is automatically ena 400 CPU thread. Per core mode is automatically enabled
488 and -a (global monitoring) is needed, requirin 401 and -a (global monitoring) is needed, requiring root rights or
489 perf.perf_event_paranoid=-1. 402 perf.perf_event_paranoid=-1.
490 403
491 Topdown uses the full Performance Monitoring U 404 Topdown uses the full Performance Monitoring Unit, and needs
492 disabling of the NMI watchdog (as root): 405 disabling of the NMI watchdog (as root):
493 echo 0 > /proc/sys/kernel/nmi_watchdog 406 echo 0 > /proc/sys/kernel/nmi_watchdog
494 for best results. Otherwise the bottlenecks ma 407 for best results. Otherwise the bottlenecks may be inconsistent
495 on workload with changing phases. 408 on workload with changing phases.
496 409
497 To interpret the results it is usually needed 410 To interpret the results it is usually needed to know on which
498 CPUs the workload runs on. If needed the CPUs 411 CPUs the workload runs on. If needed the CPUs can be forced using
499 taskset. 412 taskset.
500 413
501 --record-tpebs:: <<
502 Enable automatic sampling on Intel TPEBS retir <<
503 modifier). Without this option, perf would not <<
504 at runtime. Currently, a zero value is assigne <<
505 this option is not set. The TPEBS hardware fea <<
506 Rapids microarchitecture. This option only exi <<
507 Intel platforms with TPEBS feature. <<
508 <<
509 --td-level:: 414 --td-level::
510 Print the top-down statistics that equal the i !! 415 Print the top-down statistics that equal to or lower than the input level.
511 users to print the interested top-down metrics !! 416 It allows users to print the interested top-down metrics level instead of
512 level 1 top-down metrics. !! 417 the complete top-down metrics.
513 !! 418
514 As the higher levels gather more metrics and u !! 419 The availability of the top-down metrics level depends on the hardware. For
515 will be less accurate. By convention a metric !! 420 example, Ice Lake only supports L1 top-down metrics. The Sapphire Rapids
516 appending '_group' to it and this will increas !! 421 supports both L1 and L2 top-down metrics.
517 gathering all metrics for a level. For example <<
518 highlight 'tma_frontend_bound'. This metric ma <<
519 'tma_frontend_bound_group' with <<
520 'perf stat -M tma_frontend_bound_group...'. <<
521 422
>> 423 Default: 0 means the max level that the current hardware support.
522 Error out if the input is higher than the supp 424 Error out if the input is higher than the supported max level.
523 425
>> 426 --no-merge::
>> 427 Do not merge results from same PMUs.
>> 428
>> 429 When multiple events are created from a single event specification,
>> 430 stat will, by default, aggregate the event counts and show the result
>> 431 in a single row. This option disables that behavior and shows
>> 432 the individual events and counts.
>> 433
>> 434 Multiple events are created from a single event specification when:
>> 435 1. Prefix or glob matching is used for the PMU name.
>> 436 2. Aliases, which are listed immediately after the Kernel PMU events
>> 437 by perf list, are used.
>> 438
524 --smi-cost:: 439 --smi-cost::
525 Measure SMI cost if msr/aperf/ and msr/smi/ ev 440 Measure SMI cost if msr/aperf/ and msr/smi/ events are supported.
526 441
527 During the measurement, the /sys/device/cpu/fr 442 During the measurement, the /sys/device/cpu/freeze_on_smi will be set to
528 freeze core counters on SMI. 443 freeze core counters on SMI.
529 The aperf counter will not be effected by the 444 The aperf counter will not be effected by the setting.
530 The cost of SMI can be measured by (aperf - un 445 The cost of SMI can be measured by (aperf - unhalted core cycles).
531 446
532 In practice, the percentages of SMI cycles is 447 In practice, the percentages of SMI cycles is very useful for performance
533 oriented analysis. --metric_only will be appli 448 oriented analysis. --metric_only will be applied by default.
534 The output is SMI cycles%, equals to (aperf - 449 The output is SMI cycles%, equals to (aperf - unhalted core cycles) / aperf
535 450
536 Users who wants to get the actual value can ap 451 Users who wants to get the actual value can apply --no-metric-only.
537 452
538 --all-kernel:: 453 --all-kernel::
539 Configure all used events to run in kernel spa 454 Configure all used events to run in kernel space.
540 455
541 --all-user:: 456 --all-user::
542 Configure all used events to run in user space 457 Configure all used events to run in user space.
543 458
544 --percore-show-thread:: 459 --percore-show-thread::
545 The event modifier "percore" has supported to 460 The event modifier "percore" has supported to sum up the event counts
546 for all hardware threads in a core and show th 461 for all hardware threads in a core and show the counts per core.
547 462
548 This option with event modifier "percore" enab 463 This option with event modifier "percore" enabled also sums up the event
549 counts for all hardware threads in a core but 464 counts for all hardware threads in a core but show the sum counts per
550 hardware thread. This is essentially a replace 465 hardware thread. This is essentially a replacement for the any bit and
551 convenient for post processing. 466 convenient for post processing.
552 467
553 --summary:: 468 --summary::
554 Print summary for interval mode (-I). 469 Print summary for interval mode (-I).
555 470
556 --no-csv-summary:: <<
557 Don't print 'summary' at the first column for <<
558 This option must be used with -x and --summary <<
559 <<
560 This option can be enabled in perf config by s <<
561 'stat.no-csv-summary'. <<
562 <<
563 $ perf config stat.no-csv-summary=true <<
564 <<
565 --cputype:: <<
566 Only enable events on applying cpu with this t <<
567 (e.g. core or atom)" <<
568 <<
569 EXAMPLES 471 EXAMPLES
570 -------- 472 --------
571 473
572 $ perf stat \-- make !! 474 $ perf stat -- make
573 475
574 Performance counter stats for 'make': 476 Performance counter stats for 'make':
575 477
576 83723.452481 task-clock:u (msec) 478 83723.452481 task-clock:u (msec) # 1.004 CPUs utilized
577 0 context-switches:u 479 0 context-switches:u # 0.000 K/sec
578 0 cpu-migrations:u 480 0 cpu-migrations:u # 0.000 K/sec
579 3,228,188 page-faults:u 481 3,228,188 page-faults:u # 0.039 M/sec
580 229,570,665,834 cycles:u 482 229,570,665,834 cycles:u # 2.742 GHz
581 313,163,853,778 instructions:u 483 313,163,853,778 instructions:u # 1.36 insn per cycle
582 69,704,684,856 branches:u 484 69,704,684,856 branches:u # 832.559 M/sec
583 2,078,861,393 branch-misses:u 485 2,078,861,393 branch-misses:u # 2.98% of all branches
584 486
585 83.409183620 seconds time elapsed 487 83.409183620 seconds time elapsed
586 488
587 74.684747000 seconds user 489 74.684747000 seconds user
588 8.739217000 seconds sys 490 8.739217000 seconds sys
589 491
590 TIMINGS 492 TIMINGS
591 ------- 493 -------
592 As displayed in the example above we can displ 494 As displayed in the example above we can display 3 types of timings.
593 We always display the time the counters were e 495 We always display the time the counters were enabled/alive:
594 496
595 83.409183620 seconds time elapsed 497 83.409183620 seconds time elapsed
596 498
597 For workload sessions we also display time the 499 For workload sessions we also display time the workloads spent in
598 user/system lands: 500 user/system lands:
599 501
600 74.684747000 seconds user 502 74.684747000 seconds user
601 8.739217000 seconds sys 503 8.739217000 seconds sys
602 504
603 Those times are the very same as displayed by 505 Those times are the very same as displayed by the 'time' tool.
604 506
605 CSV FORMAT 507 CSV FORMAT
606 ---------- 508 ----------
607 509
608 With -x, perf stat is able to output a not-qui 510 With -x, perf stat is able to output a not-quite-CSV format output
609 Commas in the output are not put into "". To m 511 Commas in the output are not put into "". To make it easy to parse
610 it is recommended to use a different character 512 it is recommended to use a different character like -x \;
611 513
612 The fields are in this order: 514 The fields are in this order:
613 515
614 - optional usec time stamp in fraction 516 - optional usec time stamp in fractions of second (with -I xxx)
615 - optional CPU, core, or socket identi 517 - optional CPU, core, or socket identifier
616 - optional number of logical CPUs aggr 518 - optional number of logical CPUs aggregated
617 - counter value 519 - counter value
618 - unit of the counter value or empty 520 - unit of the counter value or empty
619 - event name 521 - event name
620 - run time of counter 522 - run time of counter
621 - percentage of measurement time the c 523 - percentage of measurement time the counter was running
622 - optional variance if multiple values 524 - optional variance if multiple values are collected with -r
623 - optional metric value 525 - optional metric value
624 - optional unit of metric 526 - optional unit of metric
625 527
626 Additional metrics may be printed with all ear 528 Additional metrics may be printed with all earlier fields being empty.
627 <<
628 include::intel-hybrid.txt[] <<
629 <<
630 JSON FORMAT <<
631 ----------- <<
632 <<
633 With -j, perf stat is able to print out a JSON <<
634 that can be used for parsing. <<
635 <<
636 - timestamp : optional usec time stamp in frac <<
637 - optional aggregate options: <<
638 - core : core identifier (with <<
639 - die : die identifier (with - <<
640 - socket : socket identifier ( <<
641 - node : node identifier (with <<
642 - thread : thread identifier ( <<
643 - counter-value : counter value <<
644 - unit : unit of the counter value or empty <<
645 - event : event name <<
646 - variance : optional variance if multiple val <<
647 - runtime : run time of counter <<
648 - metric-value : optional metric value <<
649 - metric-unit : optional unit of metric <<
650 529
651 SEE ALSO 530 SEE ALSO
652 -------- 531 --------
653 linkperf:perf-top[1], linkperf:perf-list[1] 532 linkperf:perf-top[1], linkperf:perf-list[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.