docs: admin-guide: clarify perf bench all behavior

The workload tracing guide lists a fixed set of benchmarks for
"perf bench all". This list is stale and can become outdated when
perf adds, removes, or renames benchmark collections or individual
benchmarks.

Describe "perf bench all" as running all available benchmarks in the perf
bench framework instead. Also document how to list the collections and
benchmarks available on a given system.

Signed-off-by: Cheng-Han Wu <hank20010209@gmail.com>
Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260503101429.254394-4-hank20010209@gmail.com>
This commit is contained in:
Cheng-Han Wu
2026-05-03 18:14:28 +08:00
committed by Jonathan Corbet
parent 6e13c0aac6
commit 7fb658ede0

View File

@@ -243,13 +243,21 @@ which can help mitigate performance regressions. It also acts as a common
benchmarking framework, enabling developers to easily create test cases,
integrate transparently, and use performance-rich tooling.
"perf bench all" command runs the following benchmarks:
"perf bench all" runs all available benchmarks in the perf bench
framework. The exact set of benchmarks depends on the perf version and on
the features enabled when perf was built.
* sched/messaging
* sched/pipe
* syscall/basic
* mem/memcpy
* mem/memset
To list the benchmark collections available on the current system, run::
perf bench
To list benchmarks in a collection, run::
perf bench <collection>
For example, to list the benchmarks in the mem collection, run::
perf bench mem
What is stress-ng and how do we use it?
=======================================