[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gentoo-user
Subject:    Re: [gentoo-user] "emerge koffice" seg faults! (it's =?iso-8859-15?q?getting	worse!?=)
From:       "Bruce J.A. Nourish" <kode187 () kode187 ! net>
Date:       2003-02-02 3:18:17
[Download RAW message or body]

On Saturday 01 February 2003 05:06 pm, Sundance wrote:
> >  If you get a seg fault while compiling qt or libkde because of bad
> > memory or a cpu overheat, I think you would get it when compiling
> > anything of any size.
>
> This is an often made and legitimate assumption, but it almost always
> ends up getting proven wrong.
>
> See this thread:
> http://marc.theaimsgroup.com/?l=gentoo-user&m=104334984122667&w=2

Not to be pedantic, but this thread doesn't really _proove_ anything. It 
simply demonstrates the only reliable thing about hardware sig11 
problems - randomness. The only other correspondence is with heat - and 
even that isn't always true.

I had sig11 problems compiling GCC/glibc on an LFS box, and eventually 
had to buy a huge fan to make it compile. No C++ involved. C++ 
compilation is not "more intensive": it simply requires less work. The 
processor does (roughly) the same amount of work for each (say) 
millisecond, it just does this work for fewer milliseconds. Now, what 
might be (and almost certainly is) true is that a different agorithm is 
used by the compiler, resulting in different patterns of memory access, 
shifts, arithmetic and interrupts. The C++ compiler's pattern 
apparently annoys the processor more.

-- 
Bruce J.A. Nourish <kode187@kode187.net>


--
gentoo-user@gentoo.org mailing list

[prev in list] [next in list] [prev in thread] [next in thread]