lkml.org 
[lkml]   [2002]   [Apr]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: Why HZ on i386 is 100 ?
>>>>> On Sun, 21 Apr 2002 18:00:22 +0000, Pavel Machek <pavel@suse.cz> said:

Pavel> .5% still looks like a lot to me. Good compiler optimization
Pavel> is .5% on average...

Umh, but those optimizations are interesting only if they're
cumulative, i.e., once you've got 10 of them and they make a *total*
difference of 5% (actually, I'm doubtful anyone really notices
differences of 20-30% other than for benchmarking purposes... ;-).
For me, 1% is the magic threshold. If we find real apps that get a
higher penalty than that, I'd either lower the HZ or see if we can
tune the timer tick to be within a safe margin.

No matter what, though, higher tick rate clearly incurs somewhat
higher overhead. The benefit is lower application-level response time
and finer-granularity timeouts. I assume Robert has all the
benchmarks to show that. ;-)

--david
-
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: 2005-03-22 13:25    [from the cache]
©2003-2011 Jasper Spaans