lkml.org 
[lkml]   [2007]   [Apr]   [8]   [last100]   RSS Feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
/
DateSun, 08 Apr 2007 22:12:17 -0400
FromRik van Riel <>
SubjectRe: [PATCH] Scheduler: Improving the scheduler performance.
Valdis.Kletnieks@vt.edu wrote:
> On Sat, 07 Apr 2007 23:42:20 +0600, root said:
> 
>> As we know that, linux scheduler use separate runqueue for every CPU of
>> a multiprocessor system, which having an active and an expired array.If
>> we use only one expired array, then the CPUs of a multiprocessor system
>> will be able to share their expired task via the accumulated expired
>> array,
> 
> I got this far, and the first thought that popped into my head was:
> 
> "Wow.  This might actually win on a UP or small MP (2-15 CPU).  But the
> lock contention on a big 512-CPU machoflops box is likely going to *suck*".
> 
> For that matter, my quick eyeballing of the code, although it doesn't *find*
> any race conditions, doesn't convince me there's any protection taken to make
> sure there aren't any.  Is there some subtle algorithmic trick I'm missing
> to ensure Nothing Bad Can Happen?

Lock contention is going to be the least of your worries.

Destroying CPU affinity is the big one I suspect.

-- 
Politics is the struggle between those who want to make their country
the best in the world, and those who believe it already is.  Each group
calls the other unpatriotic.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2007-04-09 02:15    [from the cache]
©2003-2008