On Sun, Jul 29, 2001 at 02:13:30PM -0700, Michael Wolfson wrote:
> Yup.  I just tried again, to double-check.  I did the CUDA reset, and my
> clocks were all messed up (symptom of the reset).  That didn't change the
> behavior of the compile.

Yeah... saw your email saying such a minute after sending that
message. Guess I should finish the mbox before replying, eh?

> What do you mean by "not quite correctly"?  My system boots, runs apps,
> gunzip -t's large files without checksum errors, and compiles a chunk of
> the kernel.  It just doesn't compile it the same way as when a 604 is
> compiling it.

I was using MacOS at the time, not NetBSD. System booted fine, but
Apple System Profiler still thought I had a 604 (I never had a 601
in the 7500, bought the motherboard second hand to replace that in
my 7200 and got the processor card separately from that) and spit
some garbage in a few of the fields. The other major sign was that
Unreal would run, but then wig out (probably some kind of
instruction error) during a game. Resetting the motherboard fixed my
problems (and I did get a bitch about the clock, if memory serves,
but was running MacOS, so it was easy to reset).

The reason I suggested this as a possibity is that I thought the gcc
binary might be wigging out the same way Unreal did under MacOS for
me, but if the NetBSD kernel's detecting the card as a G3, that
won't be the problem. (It is, right?)

-- 
       ~ g r @ eclipsed.net