Re: How do we test the IBM ieee754 math library?

["H . J . Lu" <hjl at lucon dot org>]

On Sun, Nov 25, 2001 at 09:07:53AM -0800, H . J . Lu wrote:
> On Sun, Nov 25, 2001 at 08:46:45AM -0500, Stephen L Moshier wrote:
> > 
> > > Have you compared the IBM libm on x86 against the asm version on x86?
> > > How are the results of "make check" in glibc?
> > 
> > What I checked, more specifically, was IBM libm functions running with the
> > x86 coprocessor set to double precision.  Now I see that using the
> > default linux fpu setting of 80-bit long double mode, the IBM routines
> > totally fall to pieces.
> > 
> > 
> > On x86 linux, make check doesn't run very far without coaxing.
> > It reports things like
> >     atest-exp.out:
> >     131049 failures; 0 errors; error rate 0.00%
> >     maximum error:   0b7f0f5245eea089d210d002d3fb98d57
> >     error in exp(1): 00000000000000000000000cbffc1f5e9
> > 
> > This may not be worthwhile trying to fix except to ask what
> > computer these programs were developed on or what computers they have
> > actually been tested on.  Evidently the code is intolerant of IEEE
> > flaws so there are going to be various systems on which it dosn't work
> > properly.
> 

According to

http://oss.software.ibm.com/mathlib/

The computations are bit-by-bit compatible upon porting among the
supported platforms. Currently the supported platforms/compilers
are: x86/Linux/gcc, x86/NT/[gcc| cl], Sparc/Solaris/gcc,
PowerPC/AIX[32bit | 64bit]/[gcc | xlc]. 

I will check it out, like what I did with the old IEEE math library and
the Intel IA64 math library.


H.J.