lkml.org 
[lkml]   [1996]   [May]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: i586 kernel optimizations?
Date
> 
> > I was curious as to why Rober L Krawitz's memcpy patches for the i586 had
> > not been implemented into the kernel?? I run pre2.0.8 with the patch, and
>
> Because we are trying to knock the last bugs out not featurize them just
> at the moment. After 2.0 I'd like to see his patches in.

Definitely.

>
> > Also, does anyone know of any other good patches to increase performance that
> > haven't been implemented into the kernel over the 1.3 series?? I have an old
>
> Building with the pentium gcc (the proper 2.7.2 one) is a big help looking at
> its output. It doesnt quite generate correct code always yet according to the
> docs so I've not booted this kernel ;)

Well, as I stated, I use 2.7.2p, and I have experienced no problems at all.
I usually optimize the kernel only at O3 however, because after playing with
the benchmarks O3 and O4 are about the same, but O6 makes some things faster
while some other things get slower.
I finally got my Byte benchmark up to a 22.0 (by unsetting all environment
variables). But even with out doing that, I was around a 21.6 to 21.8. For
a P133 this is pretty good, especially with an EIDE interface and 1k ext2
fs blocks. I compiled the benchmark binaries in ELF as well, because my
system is almost completely ELF. Aout is supposedly even faster.
I haven't installed the sched.c patch because I would like to
get readprofile to work on /proc/profile, but I can't get it to work.
Any help would be much appreciated, or another way to profile the kernel
that I haven't heard about.

Thanks!
hartlw@rpi.edu
Matt Hartley
http://www.rpi.edu/~hartlw



\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.034 / U:0.768 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site