lkml.org 
[lkml]   [1996]   [Dec]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subjectstrength-reduction

Samuel S Chessman wrote:

>I do not use -fno-strength-reduce as I have a patched gcc 2.7.2.

-fno-strength-reduce make code _faster_ on Intel PC's. 80x86'es
dont have very many registers, and strength-reduction makes one
less register available for the "reduced" calcuation. It might
make sense on RISC CPU's with flocks of registers, but not with
486's, Pentiums or PPros.

I've done benchmark tests with this... -fno-strength-reduce
improved scores significantly (at least 5-10% on dhrystone2).
So even if you have patched to correct the bug, it's better to
leave it in.

----
Mark Orr
markorr@intersurf.com

\
 
 \ /
  Last update: 2005-03-22 13:38    [W:0.065 / U:0.284 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site