[Freepooma-devel] Re: ReRe: [pooma-dev] SIMD

[Roman Krylov]

Hi all.
Richard,
gcc 4.0 (I have gcc version 4.0.0 20050130 (experimental)) provides 
autovectorization for c++,
but, as you had noticed it doesn't vectorize pooma loops.
Is the main reason for that that vectorizer can autovect only local & 
aligned arrays?
Maybe it would be good to have some static cache and distribute it among 
args and ret : split evaluation loop by the number of num_args+1 and 
have internal loop with that size.
The cache is filled by another loop before that inner loop is reached;
The size of the cache is determined by the user on program startup;
Sorry if I'm talking nonsense, merely I am fascinated by vectorizer - on 
primitive loops it gave 2.3 factor in performance on my P4(with sse2 as 
I think).
Roman.

Richard Guenther wrote:

> On Tue, 19 Oct 2004, Roman Krylov wrote:
>
>  
>
> > Vectorization rulez! Good feature.
> > But they say there is no vectorization support for c++ in gcc yet.
> > On gcc.gnu.org they say the next release will be 4.0, but you talk about
> > 4.1?
> > Could you give me some link about these please?
> >    
>
> Yes, the next release will be 4.0 and that features -ftree-vectorize
> which is initial support of auto-vectorization.  But the auto-vectorizer
> does not yet handle the loops produced by POOMA, so I'm betting on
> 4.1 which has not even started (though you can try the lno branch
> which features a somewhat extended auto-vectorizer that still cannot
> handle POOMA loops).
>
> So don't count on POOMA + SIMD in this or the next year.
>
> Richard.
>
> --
> Richard Guenther <richard dot guenther at uni-tuebingen dot de>
> WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/
>
>
> .
>
>