This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH][ppc] merge in powerpc version of slowpow.c
On Thu, Jan 03, 2013 at 12:14:09PM +0100, Andreas Jaeger wrote:
> >/* Compute pow as long double, which is 106 bits precision for IBM
> > long double and 112 bits for IEE754 long double format. */
>
> x86-64 uses the code as well and has 80-bit precision.
>
Oh, then I think my patch would be wrong for x86 - I was considering
only the ibm and ieee long double types and missed the fact that
x86-64 long double is just 80 bits. The constant 0x4.0-p96 will be
useless since 0x1.0 + 0x4.0-p96 will get truncated to 0x1.0, rendering
the last accuracy check useless.
Siddhesh