6 perf-record - Run a command and record its profile into perf.data
11 'perf record' [-e <EVENT> | --event=EVENT] [-l] [-a] <command>
12 'perf record' [-e <EVENT> | --event=EVENT] [-l] [-a] -- <command> [<options>]
16 This command runs a command and gathers a performance counter profile
17 from it, into perf.data - without displaying anything.
19 This file can then be inspected later on, using 'perf report'.
25 Any command you can specify in a shell.
29 Select the PMU event. Selection can be:
31 - a symbolic event name (use 'perf list' to list all events)
33 - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a
34 hexadecimal event descriptor.
36 - a symbolically formed PMU event like 'pmu/param1=0x3,param2/' where
37 'param1', 'param2', etc are defined as formats for the PMU in
38 /sys/bus/event_sources/devices/<pmu>/format/*.
40 - a symbolically formed event like 'pmu/config=M,config1=N,config3=K/'
42 where M, N, K are numbers (in decimal, hex, octal format). Acceptable
43 values for each of 'config', 'config1' and 'config2' are defined by
44 corresponding entries in /sys/bus/event_sources/devices/<pmu>/format/*
45 param1 and param2 are defined as formats for the PMU in:
46 /sys/bus/event_sources/devices/<pmu>/format/*
48 - a hardware breakpoint event in the form of '\mem:addr[/len][:access]'
49 where addr is the address in memory you want to break in.
50 Access is the memory access type (read, write, execute) it can
51 be passed as follows: '\mem:addr[:[r][w][x]]'. len is the range,
52 number of bytes from specified addr, which the breakpoint will cover.
53 If you want to profile read-write accesses in 0x1000, just set
55 If you want to profile write accesses in [0x1000~1008), just set
58 - a group of events surrounded by a pair of brace ("{event1,event2,...}").
59 Each event is separated by commas and the group should be quoted to
60 prevent the shell interpretation. You also need to use --group on
61 "perf report" to view group events together.
68 System-wide collection from all CPUs.
72 Record events on existing process ID (comma separated list).
76 Record events on existing thread ID (comma separated list).
77 This option also disables inheritance by default. Enable it by adding
82 Record events in threads owned by uid. Name or number.
86 Collect data with this RT SCHED_FIFO priority.
89 Collect data without buffering.
93 Event period to sample.
101 Child tasks do not inherit counters.
104 Profile at this frequency.
108 Number of mmap data pages (must be a power of two) or size
109 specification with appended unit character - B/K/M/G. The
110 size is rounded up to have nearest pages power of two value.
111 Also, by adding a comma, the number of mmap pages for AUX
112 area tracing can be specified.
115 Put all events in a single event group. This precedes the --event
116 option and remains only for backward compatibility. See --event.
119 Enables call-graph (stack chain/backtrace) recording.
122 Setup and enable call-graph (stack chain/backtrace) recording,
125 Allows specifying "fp" (frame pointer) or "dwarf"
126 (DWARF's CFI - Call Frame Information) or "lbr"
127 (Hardware Last Branch Record facility) as the method to collect
128 the information used to show the call graphs.
130 In some systems, where binaries are build with gcc
131 --fomit-frame-pointer, using the "fp" method will produce bogus
132 call graphs, using "dwarf", if available (perf tools linked to
133 the libunwind library) should be used instead.
134 Using the "lbr" method doesn't require any compiler options. It
135 will produce call graphs from the hardware LBR registers. The
136 main limition is that it is only available on new Intel
137 platforms, such as Haswell. It can only get user call chain. It
138 doesn't work with branch stack sampling at the same time.
142 Don't print any message, useful for scripting.
146 Be more verbose (show counter open errors, etc).
150 Record per-thread event counts. Use it with 'perf report -T' to see
155 Record the sample addresses.
159 Record the sample timestamps. Use it with 'perf report -D' to see the
160 timestamps, for instance.
164 Record the sample period.
172 Collect raw sample records from all opened counters (default for tracepoint counters).
176 Collect samples only on the list of CPUs provided. Multiple CPUs can be provided as a
177 comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2.
178 In per-thread mode with inheritance mode on (default), samples are captured only when
179 the thread executes on the designated CPUs. Default is to monitor all CPUs.
183 Do not update the buildid cache. This saves some overhead in situations
184 where the information in the perf.data file (which includes buildids)
189 monitor only in the container (cgroup) called "name". This option is available only
190 in per-cpu mode. The cgroup filesystem must be mounted. All threads belonging to
191 container "name" are monitored when they run on the monitored CPUs. Multiple cgroups
192 can be provided. Each cgroup is applied to the corresponding event, i.e., first cgroup
193 to first event, second cgroup to second event and so on. It is possible to provide
194 an empty cgroup (monitor all the time) using, e.g., -G foo,,bar. Cgroups must have
195 corresponding events, i.e., they always refer to events defined earlier on the command
200 Enable taken branch stack sampling. Any type of taken branch may be sampled.
201 This is a shortcut for --branch-filter any. See --branch-filter for more infos.
205 Enable taken branch stack sampling. Each sample captures a series of consecutive
206 taken branches. The number of branches captured with each sample depends on the
207 underlying hardware, the type of branches of interest, and the executed code.
208 It is possible to select the types of branches captured by enabling filters. The
209 following filters are defined:
211 - any: any type of branches
212 - any_call: any function call or system call
213 - any_ret: any function return or system call return
214 - ind_call: any indirect branch
215 - u: only when the branch target is at the user level
216 - k: only when the branch target is in the kernel
217 - hv: only when the target is at the hypervisor level
218 - in_tx: only when the target is in a hardware transaction
219 - no_tx: only when the target is not in a hardware transaction
220 - abort_tx: only when the target is a hardware transaction abort
221 - cond: conditional branches
224 The option requires at least one branch type among any, any_call, any_ret, ind_call, cond.
225 The privilege levels may be omitted, in which case, the privilege levels of the associated
226 event are applied to the branch filter. Both kernel (k) and hypervisor (hv) privilege
227 levels are subject to permissions. When sampling on multiple events, branch stack sampling
228 is enabled for all the sampling events. The sampled branch type is the same for all events.
229 The various filters must be specified as a comma separated list: --branch-filter any_ret,u,k
230 Note that this feature may not be available on all processors.
233 Enable weightened sampling. An additional weight is recorded per sample and can be
234 displayed with the weight and local_weight sort keys. This currently works for TSX
235 abort events and some memory events in precise mode on modern Intel CPUs.
238 Record transaction flags for transaction related events.
241 Use per-thread mmaps. By default per-cpu mmaps are created. This option
242 overrides that and uses per-thread mmaps. A side-effect of that is that
243 inheritance is automatically disabled. --per-thread is ignored with a warning
244 if combined with -a or -C options.
248 After starting the program, wait msecs before measuring. This is useful to
249 filter out the startup phase of the program, which is often very different.
253 Capture machine state (registers) at interrupt, i.e., on counter overflows for
254 each sample. List of captured registers depends on the architecture. This option
258 Record running and enabled time for read events (:S)
262 Sets the clock id to use for the various time fields in the perf_event_type
263 records. See clock_gettime(). In particular CLOCK_MONOTONIC and
264 CLOCK_MONOTONIC_RAW are supported, some events might also allow
265 CLOCK_BOOTTIME, CLOCK_REALTIME and CLOCK_TAI.
269 Select AUX area tracing Snapshot Mode. This option is valid only with an
270 AUX area tracing event. Optionally the number of bytes to capture per
271 snapshot can be specified. In Snapshot Mode, trace data is captured only when
272 signal SIGUSR2 is received.
275 When processing pre-existing threads /proc/XXX/mmap, it may take a long time,
276 because the file may be huge. A time out is needed in such cases.
277 This option sets the time out limit. The default value is 500 ms.
281 linkperf:perf-stat[1], linkperf:perf-list[1]