blob: bc5990c33dc0dd3b0e81d3fcd0a9a80c2e1a6ce6 [file] [log] [blame]
Arjan van de Ven151750c2009-09-19 13:36:04 +02001perf-timechart(1)
2=================
3
4NAME
5----
6perf-timechart - Tool to visualize total system behavior during a workload
7
8SYNOPSIS
9--------
10[verse]
Stanislav Fomichev367b3152013-11-01 20:25:50 +040011'perf timechart' [<timechart options>] {record} [<record options>]
Arjan van de Ven151750c2009-09-19 13:36:04 +020012
13DESCRIPTION
14-----------
15There are two variants of perf timechart:
16
17 'perf timechart record <command>' to record the system level events
18 of an arbitrary workload.
19
20 'perf timechart' to turn a trace into a Scalable Vector Graphics file,
21 that can be viewed with popular SVG viewers such as 'Inkscape'.
22
Stanislav Fomichev367b3152013-11-01 20:25:50 +040023TIMECHART OPTIONS
24-----------------
Arjan van de Ven151750c2009-09-19 13:36:04 +020025-o::
26--output=::
27 Select the output file (default: output.svg)
28-i::
29--input=::
Robert Richterefad1412011-12-07 10:02:54 +010030 Select the input file (default: perf.data unless stdin is a fifo)
Arjan van de Ven5094b652009-09-20 18:14:16 +020031-w::
32--width=::
33 Select the width of the SVG file (default: 1000)
Arjan van de Venbbe29872009-10-20 07:09:39 +090034-P::
Arjan van de Ven39a90a82009-09-24 15:40:13 +020035--power-only::
36 Only output the CPU power section of the diagram
Stanislav Fomichevc87097d2013-11-01 20:25:48 +040037-T::
38--tasks-only::
39 Don't output processor state transitions
Arjan van de Venbbe29872009-10-20 07:09:39 +090040-p::
41--process::
42 Select the processes to display, by name or PID
Arjan van de Ven151750c2009-09-19 13:36:04 +020043
David Ahernec5761e2010-12-09 13:27:07 -070044--symfs=<directory>::
45 Look for files with symbols relative to this directory.
Arjan van de Ven151750c2009-09-19 13:36:04 +020046
Ramkumar Ramachandrad366c532013-10-03 14:45:16 +053047EXAMPLES
48--------
49
50$ perf timechart record git pull
51
52 [ perf record: Woken up 13 times to write data ]
53 [ perf record: Captured and wrote 4.253 MB perf.data (~185801 samples) ]
54
55$ perf timechart
56
57 Written 10.2 seconds of trace to output.svg.
58
Stanislav Fomicheve57a2df2013-12-17 19:53:49 +040059Record system-wide timechart:
60
61 $ perf timechart record
62
63 then generate timechart and highlight 'gcc' tasks:
64
65 $ perf timechart --highlight gcc
66
Stanislav Fomichev54874e32013-11-01 20:25:46 +040067-n::
68--proc-num::
69 Print task info for at least given number of tasks.
Stanislav Fomichevc5079992013-12-02 18:37:36 +040070-t::
71--topology::
72 Sort CPUs according to topology.
Stanislav Fomicheve57a2df2013-12-17 19:53:49 +040073--highlight=<duration_nsecs|task_name>::
74 Highlight tasks (using different color) that run more than given
75 duration or tasks with given name. If number is given it's interpreted
76 as number of nanoseconds. If non-numeric string is given it's
77 interpreted as task name.
Stanislav Fomichev54874e32013-11-01 20:25:46 +040078
Stanislav Fomichev367b3152013-11-01 20:25:50 +040079RECORD OPTIONS
80--------------
81-P::
82--power-only::
83 Record only power-related events
84-T::
85--tasks-only::
86 Record only tasks-related events
Stanislav Fomichev6f8d67f2013-11-01 20:25:51 +040087-g::
88--callchain::
89 Do call-graph (stack chain/backtrace) recording
Stanislav Fomichev367b3152013-11-01 20:25:50 +040090
Arjan van de Ven151750c2009-09-19 13:36:04 +020091SEE ALSO
92--------
93linkperf:perf-record[1]