This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: correct rounding or not?
Dear Wilco,
> __sin used to have this:
>
> else if (k < 0x3feb6000)
> {
> res = do_sin (x, 0, &cor);
> retval = (res == res + 1.096 * cor) ? res : slow1 (x);
this would correspond to a maximum error of 0.5+0.096/4 = 0.524 ulp.
This corresponds to what I see after about 2.3e9 random tests (5.23e-1).
> >From this I estimated the worst-case ULP as < 0.55. It could well be a bit
> better in reality.
Best regards,
Paul