------- Comment #17 from dave at hiauly1 dot hia dot nrc dot ca 2007-11-03
22:49 -------
Subject: Re: FAIL: gfortran.dg/gamma_5.f90
> > Now, the hard part. The gfortran.dg/gamma_5.f90 test fails at n = 16.
>
> This works on my i686-pc-linux-gnu system, and also fails when I use
> -ffloat-store. Seems like we have a roundoff problem with normal ieee
> double precision.
>
> I'll continue to work on it...
Looking at Pugh's paper, he shows coefficients for n = 10 and
r = 10.900511. This is the same as you are using for the double
case. However, you seem to be missing coefficient d10. I believe
that Pugh's coefficients and your coefficients should differ by
the factor
2 * sqrt (e / pi) * exp (r)
due to Toth's rearrangement of the approximation. This should
provide a check for your coefficient calculation.
Dave
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33698
