[lkml]   [1996]   [Jun]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Pentium memcpy patch
I just resubscribed.  I was off the list for the past couple of weeks
(wedding & honeymoon).

From: (Christopher Hassell)
Date: Wed, 5 Jun 1996 11:39:14 -0600 (MDT)

> I'm dubious about using the pentium-memcpy patch as the author says he
> sees bit errors with the patch when operating under high load.

I'm the only one who's seen this happen (and I'm the author of the
patch). These errors went away for a while when I sped up the memory
timing (you read that right) on my motherboard, but have since
returned (possibly with the onset of warmer weather). I'm starting to
think that it's a thermal problem of some kind. Maybe when I run my
air conditioner harder they'll go away again.

There were problems with earlier versions of the patch that would
cause corruption under heavy load. This was not single bit corruption
in random spots; it was patterned corruption at page boundaries.

> Can anyone post saying if they are using the patch flawlessly under load,
> or conversely had these problems with the patch?

I've been using it cleanly since about 1.3.100. I grabbed a 2.0.7
kernel, re-applied it and turned it on.. and it still works great,
thus far that I can see.

The only problem I've had has to do with modules that tend to get
their grimey mitts into kernel functions that they
shouldn't... including the confusing memcpy. When they are
insmod'ed they cannot find anything.

Hmm. I thought I fixed that (or were some modules assuming somehow
that memcpy() was inlined?).

Outside of that (which is apparently not a lasting problem).. its
shown a real speedup! I only even have a P-75.

Robert Krawitz <>

Member of the League for Programming Freedom -- mail
Tall Clubs International -- or 1-800-521-2512

 \ /
  Last update: 2005-03-22 13:37    [W:0.115 / U:0.668 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site