This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: 100% speed improvement 3.4 -> 3.5 for ia64!! (?)
On Wed, 14 Apr 2004, Richard Guenther wrote:
> Hi!
>
> I just got the first run-time results for a gcc3.5 ia64 run with leafified
> tramp3d-v3 testcase and am impressed by the improvement compared to the
> 3.4 release candidate. 3.4 takes 7.34s/it while 3.5 finishes in an
> astonishing 4s/it. At a first glance, numerical results are within error.
Just to confirm what I already thought writing this report, testing the
same on ia32 reveals no differences if it not be a slight slowdown with
3.5 (4.42s/it with 3.4 vs. 4.46s/it with 3.5).
> Anyone has an idea what improved so much for ia64 compared to 3.4? Only
> flags I'm using are -O2 -funroll-loops -ffast-math, numerics are mostly
> FP multiplication/addition. Do I need to be concerned and investigate
> results more closely?
So if there is something architecture specific with that much of an
improvement, this may be a candidate for a backport to 3.4.1?
Richard.
--
Richard Guenther <richard dot guenther at uni-tuebingen dot de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/