| Stephane Eranian | 028f12e | 2013-01-24 16:10:38 +0100 | [diff] [blame] | 1 | perf-mem(1) |
| 2 | =========== |
| 3 | |
| 4 | NAME |
| 5 | ---- |
| 6 | perf-mem - Profile memory accesses |
| 7 | |
| 8 | SYNOPSIS |
| 9 | -------- |
| 10 | [verse] |
| 11 | 'perf mem' [<options>] (record [<command>] | report) |
| 12 | |
| 13 | DESCRIPTION |
| 14 | ----------- |
| 15 | "perf mem -t <TYPE> record" runs a command and gathers memory operation data |
| 16 | from it, into perf.data. Perf record options are accepted and are passed through. |
| 17 | |
| 18 | "perf mem -t <TYPE> report" displays the result. It invokes perf report with the |
| 19 | right set of options to display a memory access profile. |
| 20 | |
| Andi Kleen | b639409 | 2014-02-28 06:02:14 -0800 | [diff] [blame] | 21 | Note that on Intel systems the memory latency reported is the use-latency, |
| 22 | not the pure load (or store latency). Use latency includes any pipeline |
| 23 | queueing delays in addition to the memory subsystem latency. |
| 24 | |
| Stephane Eranian | 028f12e | 2013-01-24 16:10:38 +0100 | [diff] [blame] | 25 | OPTIONS |
| 26 | ------- |
| 27 | <command>...:: |
| 28 | Any command you can specify in a shell. |
| 29 | |
| 30 | -t:: |
| 31 | --type=:: |
| 32 | Select the memory operation type: load or store (default: load) |
| 33 | |
| 34 | -D:: |
| 35 | --dump-raw-samples=:: |
| 36 | Dump the raw decoded samples on the screen in a format that is easy to parse with |
| 37 | one sample per line. |
| 38 | |
| 39 | -x:: |
| 40 | --field-separator:: |
| 41 | Specify the field separator used when dump raw samples (-D option). By default, |
| 42 | The separator is the space character. |
| 43 | |
| 44 | -C:: |
| 45 | --cpu-list:: |
| 46 | Restrict dump of raw samples to those provided via this option. Note that the same |
| 47 | option can be passed in record mode. It will be interpreted the same way as perf |
| 48 | record. |
| 49 | |
| 50 | SEE ALSO |
| 51 | -------- |
| 52 | linkperf:perf-record[1], linkperf:perf-report[1] |