1perf-stat(1)
2============
3
4NAME
5----
6perf-stat - Run a command and gather performance counter statistics
7
8SYNOPSIS
9--------
10[verse]
11'perf stat' [-e <EVENT> | --event=EVENT] [-a] <command>
12'perf stat' [-e <EVENT> | --event=EVENT] [-a] -- <command> [<options>]
13'perf stat' [-e <EVENT> | --event=EVENT] [-a] record [-o file] -- <command> [<options>]
14'perf stat' report [-i file]
15
16DESCRIPTION
17-----------
18This command runs a command and gathers performance counter statistics
19from it.
20
21
22OPTIONS
23-------
24<command>...::
25	Any command you can specify in a shell.
26
27record::
28	See STAT RECORD.
29
30report::
31	See STAT REPORT.
32
33-e::
34--event=::
35	Select the PMU event. Selection can be:
36
37	- a symbolic event name (use 'perf list' to list all events)
38
39	- a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
40	  hexadecimal event descriptor.
41
42	- a symbolically formed event like 'pmu/param1=0x3,param2/' where
43	  param1 and param2 are defined as formats for the PMU in
44	  /sys/bus/event_source/devices/<pmu>/format/*
45
46	- a symbolically formed event like 'pmu/config=M,config1=N,config2=K/'
47	  where M, N, K are numbers (in decimal, hex, octal format).
48	  Acceptable values for each of 'config', 'config1' and 'config2'
49	  parameters are defined by corresponding entries in
50	  /sys/bus/event_source/devices/<pmu>/format/*
51
52-i::
53--no-inherit::
54        child tasks do not inherit counters
55-p::
56--pid=<pid>::
57        stat events on existing process id (comma separated list)
58
59-t::
60--tid=<tid>::
61        stat events on existing thread id (comma separated list)
62
63
64-a::
65--all-cpus::
66        system-wide collection from all CPUs (default if no target is specified)
67
68-c::
69--scale::
70	scale/normalize counter values
71
72-d::
73--detailed::
74	print more detailed statistics, can be specified up to 3 times
75
76	   -d:          detailed events, L1 and LLC data cache
77        -d -d:     more detailed events, dTLB and iTLB events
78     -d -d -d:     very detailed events, adding prefetch events
79
80-r::
81--repeat=<n>::
82	repeat command and print average + stddev (max: 100). 0 means forever.
83
84-B::
85--big-num::
86        print large numbers with thousands' separators according to locale
87
88-C::
89--cpu=::
90Count only on the list of CPUs provided. Multiple CPUs can be provided as a
91comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.
92In per-thread mode, this option is ignored. The -a option is still necessary
93to activate system-wide monitoring. Default is to count on all CPUs.
94
95-A::
96--no-aggr::
97Do not aggregate counts across all monitored CPUs.
98
99-n::
100--null::
101        null run - don't start any counters
102
103-v::
104--verbose::
105        be more verbose (show counter open errors, etc)
106
107-x SEP::
108--field-separator SEP::
109print counts using a CSV-style output to make it easy to import directly into
110spreadsheets. Columns are separated by the string specified in SEP.
111
112-G name::
113--cgroup name::
114monitor only in the container (cgroup) called "name". This option is available only
115in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to
116container "name" are monitored when they run on the monitored CPUs. Multiple cgroups
117can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup
118to first event, second cgroup to second event and so on. It is possible to provide
119an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have
120corresponding events, i.e., they always refer to events defined earlier on the command
121line.
122
123-o file::
124--output file::
125Print the output into the designated file.
126
127--append::
128Append to the output file designated with the -o option. Ignored if -o is not specified.
129
130--log-fd::
131
132Log output to fd, instead of stderr.  Complementary to --output, and mutually exclusive
133with it.  --append may be used here.  Examples:
134     3>results  perf stat --log-fd 3          -- $cmd
135     3>>results perf stat --log-fd 3 --append -- $cmd
136
137--pre::
138--post::
139	Pre and post measurement hooks, e.g.:
140
141perf stat --repeat 10 --null --sync --pre 'make -s O=defconfig-build/clean' -- make -s -j64 O=defconfig-build/ bzImage
142
143-I msecs::
144--interval-print msecs::
145Print count deltas every N milliseconds (minimum: 10ms)
146The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
147	example: 'perf stat -I 1000 -e cycles -a sleep 5'
148
149--metric-only::
150Only print computed metrics. Print them in a single line.
151Don't show any raw values. Not supported with --per-thread.
152
153--per-socket::
154Aggregate counts per processor socket for system-wide mode measurements.  This
155is a useful mode to detect imbalance between sockets.  To enable this mode,
156use --per-socket in addition to -a. (system-wide).  The output includes the
157socket number and the number of online processors on that socket. This is
158useful to gauge the amount of aggregation.
159
160--per-core::
161Aggregate counts per physical processor for system-wide mode measurements.  This
162is a useful mode to detect imbalance between physical cores.  To enable this mode,
163use --per-core in addition to -a. (system-wide).  The output includes the
164core number and the number of online logical processors on that physical processor.
165
166--per-thread::
167Aggregate counts per monitored threads, when monitoring threads (-t option)
168or processes (-p option).
169
170-D msecs::
171--delay msecs::
172After starting the program, wait msecs before measuring. This is useful to
173filter out the startup phase of the program, which is often very different.
174
175-T::
176--transaction::
177
178Print statistics of transactional execution if supported.
179
180STAT RECORD
181-----------
182Stores stat data into perf data file.
183
184-o file::
185--output file::
186Output file name.
187
188STAT REPORT
189-----------
190Reads and reports stat data from perf data file.
191
192-i file::
193--input file::
194Input file name.
195
196--per-socket::
197Aggregate counts per processor socket for system-wide mode measurements.
198
199--per-core::
200Aggregate counts per physical processor for system-wide mode measurements.
201
202-A::
203--no-aggr::
204Do not aggregate counts across all monitored CPUs.
205
206--topdown::
207Print top down level 1 metrics if supported by the CPU. This allows to
208determine bottle necks in the CPU pipeline for CPU bound workloads,
209by breaking the cycles consumed down into frontend bound, backend bound,
210bad speculation and retiring.
211
212Frontend bound means that the CPU cannot fetch and decode instructions fast
213enough. Backend bound means that computation or memory access is the bottle
214neck. Bad Speculation means that the CPU wasted cycles due to branch
215mispredictions and similar issues. Retiring means that the CPU computed without
216an apparently bottleneck. The bottleneck is only the real bottleneck
217if the workload is actually bound by the CPU and not by something else.
218
219For best results it is usually a good idea to use it with interval
220mode like -I 1000, as the bottleneck of workloads can change often.
221
222The top down metrics are collected per core instead of per
223CPU thread. Per core mode is automatically enabled
224and -a (global monitoring) is needed, requiring root rights or
225perf.perf_event_paranoid=-1.
226
227Topdown uses the full Performance Monitoring Unit, and needs
228disabling of the NMI watchdog (as root):
229echo 0 > /proc/sys/kernel/nmi_watchdog
230for best results. Otherwise the bottlenecks may be inconsistent
231on workload with changing phases.
232
233This enables --metric-only, unless overriden with --no-metric-only.
234
235To interpret the results it is usually needed to know on which
236CPUs the workload runs on. If needed the CPUs can be forced using
237taskset.
238
239--no-merge::
240Do not merge results from same PMUs.
241
242--smi-cost::
243Measure SMI cost if msr/aperf/ and msr/smi/ events are supported.
244
245During the measurement, the /sys/device/cpu/freeze_on_smi will be set to
246freeze core counters on SMI.
247The aperf counter will not be effected by the setting.
248The cost of SMI can be measured by (aperf - unhalted core cycles).
249
250In practice, the percentages of SMI cycles is very useful for performance
251oriented analysis. --metric_only will be applied by default.
252The output is SMI cycles%, equals to (aperf - unhalted core cycles) / aperf
253
254Users who wants to get the actual value can apply --no-metric-only.
255
256EXAMPLES
257--------
258
259$ perf stat -- make -j
260
261 Performance counter stats for 'make -j':
262
263    8117.370256  task clock ticks     #      11.281 CPU utilization factor
264            678  context switches     #       0.000 M/sec
265            133  CPU migrations       #       0.000 M/sec
266         235724  pagefaults           #       0.029 M/sec
267    24821162526  CPU cycles           #    3057.784 M/sec
268    18687303457  instructions         #    2302.138 M/sec
269      172158895  cache references     #      21.209 M/sec
270       27075259  cache misses         #       3.335 M/sec
271
272 Wall-clock time elapsed:   719.554352 msecs
273
274CSV FORMAT
275----------
276
277With -x, perf stat is able to output a not-quite-CSV format output
278Commas in the output are not put into "". To make it easy to parse
279it is recommended to use a different character like -x \;
280
281The fields are in this order:
282
283	- optional usec time stamp in fractions of second (with -I xxx)
284	- optional CPU, core, or socket identifier
285	- optional number of logical CPUs aggregated
286	- counter value
287	- unit of the counter value or empty
288	- event name
289	- run time of counter
290	- percentage of measurement time the counter was running
291	- optional variance if multiple values are collected with -r
292	- optional metric value
293	- optional unit of metric
294
295Additional metrics may be printed with all earlier fields being empty.
296
297SEE ALSO
298--------
299linkperf:perf-top[1], linkperf:perf-list[1]
300