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