> It's odd, because I can use only one of the two DIMMs, or both,
   > interchanging in different combinations, and it's always gcc dying with the
   > same strange errors every time, to the letter, on the same file.  Nothing
   > else is dying. 

   I have only one DIMM (128MB), but I saw similar behavior on the
   i386 (-current).  If I ran the distributed.net rc5des client and
   tried to build a kernel, cc1 would (like clockwork) drop core on
   (I think) cd9660_vnops.c.  I tried changing the BIOS settings for
   memory speed, etc., around to no avail.  If I stopped the client,
   the kernel compile proceeded.  I think I mentioned this in mail at
   some point, but I never qualified it enough to make it into a bug
   report that I thought might get looked at.

   -allen

   -- 
		       Allen Briggs - briggs@ninthwonder.com



Eggg....  I have noticed this also.  I have a 233MHZ AMD system with one
64MB SDRAM for its main memory that can not run the rc5des client without
random core dumps followed by a panic soon after.  For me, it would often
panic in the FFS code.  [This was noticed on 1.3E]


Brad Spencer - brad@anduin.eldar.org   http://anduin.eldar.org
[finger brad@anduin.eldar.org for PGP public key]