[lkml]   [2005]   [Jul]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 3/3] Use conditional
Hi James,

[added to Cc]

On Sun, Jul 03, 2005 at 12:51:20PM -0400, James Morris wrote:
> On Sun, 3 Jul 2005, Kurt Garloff wrote:
> > capablities is used. These are not called via indirect calls but
> > called as hardcoded calls and might thus be inlined; the price for
> > this is a conditional -- benchmarks done by hp showed this to be
> > beneficial (on ia64).
> What about on i386, x86_64 or ppc64?

We tested on i386 as well at the time, and it looked like a tiny
improvement. But doing the statistics, it was in the noise.
I have no numbers for x86_64 or ppc64.

If you have reason to believe that there could be regressions, we
should indeed do the benchmarks.

Sidenote: The patches 1 -- 2b alone still make sense, so I would
vote not for delaying their inclusion until we can collect numbers
for all arches we care about to take a decision on patch 3.

Kurt Garloff, Director SUSE Labs, Novell Inc.
[unhandled content-type:application/pgp-signature]
 \ /
  Last update: 2005-07-03 23:21    [W:0.066 / U:1.056 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site