Re: Alternative implementation of the generic __ffs

[Alexander van Heukelum]

On Sat, 19 Apr 2008 11:17:01 -0700, "Joe Perches" <joe@xxxxxxxxxxx>
said:
> On Sat, 2008-04-19 at 14:10 +0200, Alexander van Heukelum wrote:
> > I've added that to the benchmark, which you can now find here:
> > http://heukelum.fastmail.fm/ffs/.

Thanks! Added the version you sent to the program and added the results
of the ARM processor to the page.

More ideas welcome ;).

> retested on arm:
> 
> $ gcc -Os -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3170 tics,  8326 tics
> New:            4214 tics,  8793 tics
> Smallest:       4023 tics,  7733 tics
> Small const:    3442 tics,  6188 tics
> Empty loop:     1517 tics,  2243 tics
> 
> $ gcc -O2 -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3172 tics,  7832 tics
> New:            4805 tics,  8790 tics
> Smallest:       4405 tics,  7154 tics
> Small const:    3442 tics,  5612 tics
> Empty loop:     1516 tics,  2145 tics
> 
> $ gcc -O3 -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3080 tics,  7709 tics
> New:            4723 tics,  8656 tics
> Smallest:       4333 tics,  7121 tics
> Small const:    3379 tics,  5483 tics
> Empty loop:     1447 tics,  2016 tics
> 
> >  Testing the same with 
> > "return x4 + x3 + x2 + x1 + x0;" as the last line would be
> > interesting too.
> 
> Adding is slower:
> 
> $ gcc -Os -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3152 tics,  8310 tics
> New:            4214 tics,  8789 tics
> Smallest:       4024 tics,  7737 tics
> Small const:    3538 tics,  6295 tics
> Empty loop:     1517 tics,  2243 tics
> 
> $ gcc -O2 -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3184 tics,  7849 tics
> New:            4790 tics,  8814 tics
> Smallest:       4406 tics,  7161 tics
> Small const:    3538 tics,  5806 tics
> Empty loop:     1521 tics,  2153 tics
> 
> $ gcc -O3 -fomit-frame-pointer ffs.c
> $ ./a.out
> Original:       3091 tics,  7694 tics
> New:            4718 tics,  8656 tics
> Smallest:       4333 tics,  7124 tics
> Small const:    3467 tics,  5687 tics
> Empty loop:     1445 tics,  2066 tics
> 
> 
-- 
  Alexander van Heukelum
  heukelum@xxxxxxxxxxx

-- 
http://www.fastmail.fm - And now for something completely different?

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/