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