[PATCH][fortran] Make sincos, cexpi and cbrt available

[Timothy C Prince]

-----Original Message-----
From: Tobias Burnus <burnus@net-b.de>
To: Richard Guenther <richard.guenther@gmail.com>
Date: Sat, 20 Jan 2007 00:31:20 +0100
Subject: Re: [PATCH][fortran] Make sincos, cexpi and cbrt available

Hi Richard,

Richard Guenther wrote:
>> Now, as proud owner of sincos, I tried to optimize a simple program (cf.
>> PR 30038); however, it does not use sincos(). Any ideas what goes wrong
>> here? [Same with Polyhedron's fatigue: No sincos!]
>
> Well, the sincos patch is not yet applied ;)

As said: I lost track which patch does what. I now applied:
http://gcc.gnu.org/ml/gcc-patches/2006-12/msg01737.html

With that, the run time of fatigue decreases from 19.7s to 15.8s (~20%).

>> (Side note: For fatigue, the unmodified gfortran which uses sinf/cosf
>> takes 19.7s, ifort which uses sincos needs 11.8s [ifort 10: 11.0s]
>> and sunf95 needs only 8.4s and uses sin/cos.

___________________________________
ifort uses a short vector sincos.  It's good that non-vector sincos accomplished half of the gain, where ifort gets 4 sincos results per function call.
Tim Prince