lkml.org 
[lkml]   [2011]   [Nov]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [tip:perf/urgent] perf/x86: Enable raw event access to Intel offcore events
From
Date
On Mon, 2011-11-21 at 22:45 +0100, Stephane Eranian wrote:
> On Mon, Nov 21, 2011 at 10:42 PM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> > On Mon, 2011-11-21 at 22:39 +0100, Stephane Eranian wrote:
> >> To solve this problem (in the general case), you need to know which
> >> counter is taken (or required)
> >> by ALL pinned system-wide events across all CPUs. Once you've
> >> constructed the bitmap, you can
> >> use it as the basis (used_mask) to try and schedule the group events.
> >>
> > Its worse, you also need to consider all task-pinned events if you want
> > to be complete. That's an even worse problem.
> >
> But that you know because by the time you do the simulation you already
> know the task the events are attached to. You can get to the list of events
> already attached to the task. Again this is a best effort, it could well be
> that someone else attaches pinned task events just after the simulation.

In the PAPI case yes, but what if you want to create a CPU-wide group,
then you need to consider all task-pinned events that could possibly
land on your CPU. Not fun.




\
 
 \ /
  Last update: 2011-11-21 22:51    [W:0.077 / U:0.264 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site