On Wed, 30 Mar 2005, Aaron Cordova wrote:

> Sorry for the confusion, I was exhausted when I initially wrote this
> email. When I start compiling using a newly built gcc it crashes (seg
> faults) a whole lot. But the errors don't happen at the same point every
> time. And always occur after varied times. If I use the gcc that comes
> with my host system (FC3) I don't get this problem.
>
 FWIW, compiling a big package (gcc, glibc, X) is the worst test of
hardware.  As well as ventilation, you may need to consider the voltages
set up on the motherboard - unlikely, but maybe you fiddled with the
bios settings in the past.  On my insane powerpc (now dead, from
repeated overheating), gcc-3.3.2 was the last "easy" compiler.
gcc-3.3.3 provoked random segfaults which I eventually attributed to too
little voltage (controlled by dip-switches on that mobo, and they were
set very conservatively).  Every gcc after 3.3.3 caused the cpu to shut
down from overheating, I sometimes thought it was a battle between me
trying to find better heatsinks that would fit, and the gcc hackers
trying to squeeze more from the hardware ;)

 As has been said, check for dust in the fan and heatsink (or if it's
old enough, check the cpu fan still works), then think about upgrading
the cooling - even small changes to the way everything is laid out, or
tidying the cabling, can sometimes help.  Also, check your power supply
is adequately specified for whichever rail your cpu draws its current
from.

 If this is happening on even small packages, you've got a serious
problem - check the fan first.  Mostly, the problem only surfaces on big
packages, and you can just repeat 'make', perhaps after letting the cpu
idle to cool down, until it finishes.

Ken
 --
 das eine Mal als Trag�die, das andere Mal als Farce

--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page

Reply via email to