[NTLUG:Discuss] AMD K6 200 vs Pentium MMX 200

[Steve Baker]

"Gregory A. Edwards" wrote:
> 
> On Thu, 23 Dec 1999, you wrote:
> >
> > THere is a hardware bug in at least some AMD K6-200's.  This bug
> > only seems to show up under Linux - and had only been seen to affect
> > the GNU C/C++ compilers - and only then if you have more than 32Mb RAM.

> Is this a problem with the AMD running the compiler or the AMD running the
> compiled code?

The problem has only been observed "in nature" in the compiler itself - although
people have contrived test code to deliberately trigger the bug.

I have certainly never seen a problem with other programs (whether
compiled by GNU compilers or not) - but it's certainly not impossible
for other programs to trigger the problem.

> I have an AMD K6 233 that is currently a Win9x workstation only
> that I'm about to setup as a dual boot.  My Cyrix 6x86 200 has shown no
> problems using the 586 (?) option for kernel compiles.  I have no intentions to
> do any upgrades of cpu on the Win9x box to add the dual, it's a 64M machine.
> This machine will also be a workstation for users and no development will be
> done on it so other than the kernel I have no need to do any complies.  I
> assume the out-of-the box kernel (RH6.1) does not have any problems?

No - in my experience, the *ONLY* problem is that sometimes when you are running
the compiler, it'll bomb out with an 'internal error' message. You re-run it
and it works the second time (nearly always)...or the third time...or...

Everything else seems to work just fine....but that's only a matter of luck
I guess.  This must be an EXTREMELY unlikely set of circumstances - what's
suprising is that I've tried maybe five different revs of the C/C++ compiler
suite and they ALL exhibit the problem!

-- 
Steve Baker                  http://web2.airmail.net/sjbaker1
sjbaker1 at airmail.net (home)  http://www.woodsoup.org/~sbaker
sjbaker at hti.com      (work)