lkml.org 
[lkml]   [2014]   [Oct]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH v2 0/12] perf/x86: implement HT leak workaround for SNB/IVB/HSW
From
On Thu, Oct 23, 2014 at 10:53 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> On Wed, Oct 22, 2014 at 11:04:31PM +0200, Stephane Eranian wrote:
>> Here is a simple case:
>> Limiting each HT to only 2 counters, can be any, 2 out of 4 possible.
>>
>> HT0: you measure a MEM* in ctr2, it is started first, and it keeps running
>> HT1: you measure PREC_DIST with PEBS (it requires ctr2)
>>
>> HT0 is measuring a corrupting event on ctr2, this prevents ctr2 on HT1
>> from being used.
>> HT1 is starved, it cannot measure PREC_DIST
>>
>> Yes you have a quota of 2 out of 4 counters.
>>
>> The quota dynamic or static can help mitigate the starvation. The only
>> way to eliminate
>> it is to force multiplexing even though you are using fewer counters
>> than actually avail.
>
> Ah yes, the very narrowly constrained events. Those suck indeed. And I
> imagine rotation might not even help here -- rotation doesn't guarantee
> SMT1 will try and schedule before SMT0, in fact there are setups
> (staggered tick) where its almost guaranteed not to.
>
> Still I suppose for 'normal' event its a much better state, SMT1 can
> always schedule some events.

Yes, I agree with you. The soft partition helps. I will add that in V3.
Thanks.


\
 
 \ /
  Last update: 2014-10-23 11:21    [W:1.499 / U:0.012 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site