Messages in this thread | | | From | Namhyung Kim <> | Date | Thu, 16 Apr 2020 23:19:19 +0900 | Subject | Re: [PATCH v5 0/3] perf synthetic events |
| |
Hi Ian,
On Wed, Apr 15, 2020 at 2:40 PM Ian Rogers <irogers@google.com> wrote: > > Add a multi-threaded version of the internals synthesize benchmark. It > attempts to compute a time per event synthesized, but as it is reading > /proc there are issues: > - permissions if not run as root > - "random" nature of /proc > > By default the benchmark is disabled but can be enabled with a > flag. It has been useful in gauging the value of multi-threaded > improvements not included here as their value appears minimal. > > The patch set includes 2 patches that improve synthesis performance > and updates the benchmark numbers: > https://lore.kernel.org/lkml/20200411064248.247530-1-irogers@google.com/ > > v4 added a missing test file > v3 improved documenation, return values and added testing to the io framework > following feedback from namhyung@kernel.org. > v2 addressed single threaded synthesize benchmark issues from jolsa@redhat.com > https://lore.kernel.org/lkml/20200402154357.107873-1-irogers@google.com/ > > Ian Rogers (3): > perf bench: add a multi-threaded synthesize benchmark > tools api: add a lightweight buffered reading api > perf synthetic events: Remove use of sscanf from /proc reading
Acked-by: Namhyung Kim <namhyung@kernel.org>
Thanks Namhyung
| |