perf-stat.txt 8.8 KB
Newer Older
I
Ingo Molnar 已提交
1
perf-stat(1)
2
============
I
Ingo Molnar 已提交
3 4 5 6 7 8 9 10

NAME
----
perf-stat - Run a command and gather performance counter statistics

SYNOPSIS
--------
[verse]
S
Shawn Bohrer 已提交
11 12
'perf stat' [-e <EVENT> | --event=EVENT] [-a] <command>
'perf stat' [-e <EVENT> | --event=EVENT] [-a] -- <command> [<options>]
J
Jiri Olsa 已提交
13
'perf stat' [-e <EVENT> | --event=EVENT] [-a] record [-o file] -- <command> [<options>]
J
Jiri Olsa 已提交
14
'perf stat' report [-i file]
I
Ingo Molnar 已提交
15 16 17 18 19 20 21 22 23 24 25 26

DESCRIPTION
-----------
This command runs a command and gathers performance counter statistics
from it.


OPTIONS
-------
<command>...::
	Any command you can specify in a shell.

J
Jiri Olsa 已提交
27 28
record::
	See STAT RECORD.
I
Ingo Molnar 已提交
29

J
Jiri Olsa 已提交
30 31 32
report::
	See STAT REPORT.

I
Ingo Molnar 已提交
33 34
-e::
--event=::
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
	Select the PMU event. Selection can be:

	- a symbolic event name (use 'perf list' to list all events)

	- a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
	  hexadecimal event descriptor.

	- a symbolically formed event like 'pmu/param1=0x3,param2/' where
	  param1 and param2 are defined as formats for the PMU in
	  /sys/bus/event_sources/devices/<pmu>/format/*

	- a symbolically formed event like 'pmu/config=M,config1=N,config2=K/'
	  where M, N, K are numbers (in decimal, hex, octal format).
	  Acceptable values for each of 'config', 'config1' and 'config2'
	  parameters are defined by corresponding entries in
	  /sys/bus/event_sources/devices/<pmu>/format/*
I
Ingo Molnar 已提交
51

I
Ingo Molnar 已提交
52
-i::
53 54
--no-inherit::
        child tasks do not inherit counters
I
Ingo Molnar 已提交
55 56
-p::
--pid=<pid>::
57
        stat events on existing process id (comma separated list)
S
Shawn Bohrer 已提交
58 59 60

-t::
--tid=<tid>::
61
        stat events on existing thread id (comma separated list)
S
Shawn Bohrer 已提交
62

I
Ingo Molnar 已提交
63

I
Ingo Molnar 已提交
64
-a::
S
Shawn Bohrer 已提交
65
--all-cpus::
J
Jiri Olsa 已提交
66
        system-wide collection from all CPUs (default if no target is specified)
I
Ingo Molnar 已提交
67

68
-c::
S
Shawn Bohrer 已提交
69 70 71
--scale::
	scale/normalize counter values

72 73 74 75 76 77 78 79
-d::
--detailed::
	print more detailed statistics, can be specified up to 3 times

	   -d:          detailed events, L1 and LLC data cache
        -d -d:     more detailed events, dTLB and iTLB events
     -d -d -d:     very detailed events, adding prefetch events

S
Shawn Bohrer 已提交
80 81
-r::
--repeat=<n>::
82
	repeat command and print average + stddev (max: 100). 0 means forever.
I
Ingo Molnar 已提交
83

84
-B::
S
Shawn Bohrer 已提交
85
--big-num::
86 87
        print large numbers with thousands' separators according to locale

88 89
-C::
--cpu=::
S
Shawn Bohrer 已提交
90 91
Count only on the list of CPUs provided. Multiple CPUs can be provided as a
comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.
92 93 94
In per-thread mode, this option is ignored. The -a option is still necessary
to activate system-wide monitoring. Default is to count on all CPUs.

95 96 97 98 99
-A::
--no-aggr::
Do not aggregate counts across all monitored CPUs in system-wide mode (-a).
This option is only valid in system-wide mode.

S
Shawn Bohrer 已提交
100 101 102 103 104 105 106 107
-n::
--null::
        null run - don't start any counters

-v::
--verbose::
        be more verbose (show counter open errors, etc)

S
Stephane Eranian 已提交
108 109 110 111 112
-x SEP::
--field-separator SEP::
print counts using a CSV-style output to make it easy to import directly into
spreadsheets. Columns are separated by the string specified in SEP.

S
Stephane Eranian 已提交
113 114 115 116 117 118 119 120 121 122 123
-G name::
--cgroup name::
monitor only in the container (cgroup) called "name". This option is available only
in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to
container "name" are monitored when they run on the monitored CPUs. Multiple cgroups
can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup
to first event, second cgroup to second event and so on. It is possible to provide
an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have
corresponding events, i.e., they always refer to events defined earlier on the command
line.

124
-o file::
125
--output file::
126 127 128 129 130
Print the output into the designated file.

--append::
Append to the output file designated with the -o option. Ignored if -o is not specified.

131 132 133 134 135 136 137
--log-fd::

Log output to fd, instead of stderr.  Complementary to --output, and mutually exclusive
with it.  --append may be used here.  Examples:
     3>results  perf stat --log-fd 3          -- $cmd
     3>>results perf stat --log-fd 3 --append -- $cmd

138 139 140 141 142
--pre::
--post::
	Pre and post measurement hooks, e.g.:

perf stat --repeat 10 --null --sync --pre 'make -s O=defconfig-build/clean' -- make -s -j64 O=defconfig-build/ bzImage
143

144 145
-I msecs::
--interval-print msecs::
146 147 148
Print count deltas every N milliseconds (minimum: 10ms)
The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
	example: 'perf stat -I 1000 -e cycles -a sleep 5'
149

150 151
--metric-only::
Only print computed metrics. Print them in a single line.
152
Don't show any raw values. Not supported with --per-thread.
153

154
--per-socket::
155 156
Aggregate counts per processor socket for system-wide mode measurements.  This
is a useful mode to detect imbalance between sockets.  To enable this mode,
157
use --per-socket in addition to -a. (system-wide).  The output includes the
158 159 160
socket number and the number of online processors on that socket. This is
useful to gauge the amount of aggregation.

161 162 163 164 165 166
--per-core::
Aggregate counts per physical processor for system-wide mode measurements.  This
is a useful mode to detect imbalance between physical cores.  To enable this mode,
use --per-core in addition to -a. (system-wide).  The output includes the
core number and the number of online logical processors on that physical processor.

167 168 169 170
--per-thread::
Aggregate counts per monitored threads, when monitoring threads (-t option)
or processes (-p option).

171
-D msecs::
172
--delay msecs::
173 174 175
After starting the program, wait msecs before measuring. This is useful to
filter out the startup phase of the program, which is often very different.

176 177 178 179 180
-T::
--transaction::

Print statistics of transactional execution if supported.

J
Jiri Olsa 已提交
181 182 183 184 185 186 187 188
STAT RECORD
-----------
Stores stat data into perf data file.

-o file::
--output file::
Output file name.

J
Jiri Olsa 已提交
189 190 191 192 193 194 195 196
STAT REPORT
-----------
Reads and reports stat data from perf data file.

-i file::
--input file::
Input file name.

197 198 199 200 201 202 203 204 205 206
--per-socket::
Aggregate counts per processor socket for system-wide mode measurements.

--per-core::
Aggregate counts per physical processor for system-wide mode measurements.

-A::
--no-aggr::
Do not aggregate counts across all monitored CPUs.

207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238
--topdown::
Print top down level 1 metrics if supported by the CPU. This allows to
determine bottle necks in the CPU pipeline for CPU bound workloads,
by breaking the cycles consumed down into frontend bound, backend bound,
bad speculation and retiring.

Frontend bound means that the CPU cannot fetch and decode instructions fast
enough. Backend bound means that computation or memory access is the bottle
neck. Bad Speculation means that the CPU wasted cycles due to branch
mispredictions and similar issues. Retiring means that the CPU computed without
an apparently bottleneck. The bottleneck is only the real bottleneck
if the workload is actually bound by the CPU and not by something else.

For best results it is usually a good idea to use it with interval
mode like -I 1000, as the bottleneck of workloads can change often.

The top down metrics are collected per core instead of per
CPU thread. Per core mode is automatically enabled
and -a (global monitoring) is needed, requiring root rights or
perf.perf_event_paranoid=-1.

Topdown uses the full Performance Monitoring Unit, and needs
disabling of the NMI watchdog (as root):
echo 0 > /proc/sys/kernel/nmi_watchdog
for best results. Otherwise the bottlenecks may be inconsistent
on workload with changing phases.

This enables --metric-only, unless overriden with --no-metric-only.

To interpret the results it is usually needed to know on which
CPUs the workload runs on. If needed the CPUs can be forced using
taskset.
J
Jiri Olsa 已提交
239

I
Ingo Molnar 已提交
240 241 242
EXAMPLES
--------

I
Ingo Molnar 已提交
243
$ perf stat -- make -j
I
Ingo Molnar 已提交
244

I
Ingo Molnar 已提交
245
 Performance counter stats for 'make -j':
I
Ingo Molnar 已提交
246

I
Ingo Molnar 已提交
247 248 249 250 251 252 253 254
    8117.370256  task clock ticks     #      11.281 CPU utilization factor
            678  context switches     #       0.000 M/sec
            133  CPU migrations       #       0.000 M/sec
         235724  pagefaults           #       0.029 M/sec
    24821162526  CPU cycles           #    3057.784 M/sec
    18687303457  instructions         #    2302.138 M/sec
      172158895  cache references     #      21.209 M/sec
       27075259  cache misses         #       3.335 M/sec
I
Ingo Molnar 已提交
255

I
Ingo Molnar 已提交
256
 Wall-clock time elapsed:   719.554352 msecs
I
Ingo Molnar 已提交
257

258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
CSV FORMAT
----------

With -x, perf stat is able to output a not-quite-CSV format output
Commas in the output are not put into "". To make it easy to parse
it is recommended to use a different character like -x \;

The fields are in this order:

	- optional usec time stamp in fractions of second (with -I xxx)
	- optional CPU, core, or socket identifier
	- optional number of logical CPUs aggregated
	- counter value
	- unit of the counter value or empty
	- event name
	- run time of counter
	- percentage of measurement time the counter was running
	- optional variance if multiple values are collected with -r
	- optional metric value
	- optional unit of metric

Additional metrics may be printed with all earlier fields being empty.

I
Ingo Molnar 已提交
281 282
SEE ALSO
--------
283
linkperf:perf-top[1], linkperf:perf-list[1]