[MLton] Re: Hppa

Matthew Fluet fluet@cs.cornell.edu
Sat, 18 Dec 2004 00:01:19 -0500 (EST)


> On Sat, Dec 18, 2004 at 02:01:24AM +0100, Wesley W. Terpstra wrote:
> > I'm re-running the regressions now; it takes forever... I'll post the log
> > here so you guys can decide if the 'real' test counts as a regression.
>
> I've attached a log of the regression tests.
> I am a bit disturbed by how only half of the digits in 'real' agree.

Since everything else is working fine, I wouldn't be too worried.  I see
about 8000 numbers in real.ok, of which your log is differing in 16.  We
know that libm occassionally has different behavior on different
platforms.  Also, since the x86 computes at 80-bits, that might account
for differences.