[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180108211049.bbtq7uvaec3hhjyz@two.firstfloor.org>
Date: Mon, 8 Jan 2018 13:10:50 -0800
From: Andi Kleen <andi@...stfloor.org>
To: Jiri Olsa <jolsa@...hat.com>
Cc: Julia Lawall <julia.lawall@...6.fr>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Namhyung Kim <namhyung@...nel.org>,
linux-kernel@...r.kernel.org, Andi Kleen <andi@...stfloor.org>
Subject: Re: perf: documentation for cycles_no_execute event
> > Many of the x86 pipeline.json files have the brief description "Total
> > execution stalls" for both CYCLE_ACTIVITY.CYCLES_NO_EXECUTE and
> > CYCLE_ACTIVITY.STALLS_TOTAL. Should the case for
> > CYCLE_ACTIVITY.CYCLES_NO_EXECUTE have a brief description that mentions
> > cycles? Some of the files do have a public description that mentions
> > cycles, eg broadwellde/pipeline.json has "Counts number of cycles nothing
> > is executed on any execution port.", but others (eg ivytown/pipeline.json)
> > just have "Total execution stalls." twice. Maybe "Cycles where nothing is
> > executed" would be ok for CYCLE_ACTIVITY.CYCLES_NO_EXECUTE?
Yes it's all in cycles
Will clarify in a future event list update.
-Andi
Powered by blists - more mailing lists