|
From: Tom H. <to...@co...> - 2006-04-13 12:58:25
|
In message <200...@ac...>
Julian Seward <js...@ac...> wrote:
> Hmm, looks like some borkage happened (also on dellow & aston).
> Tom, could you have a look?
Mostly just that the external debuginfo commit caused it to start
picking up a source file and line number for the "(below main)" line
in various stack traces. I've patched filter_libc to fix it.
The interesting failure I noticed (and this isn't a change) is
actually memcheck/tests/x86/sse1_memory which is giving different
output for most of the rcpss/rcpps/rsqrtss/rsqrtps tests on most
of my machines (both 32 and 64 bit).
The only one where it matches is ford (FC4 on x86) which is also the
only Intel chip machine (a P4) while all the others are AMD based
machines (Athlon XP/MP/64).
I assume that AMD have a subtly different algorithm for computing
the reciprocals and square roots that gives rise to changes in the
low order bits of the mantissa or something - hard to tell from the
hex dump of the floating point number.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|