| 1 |
I agree with Ellotheth that it seems like there's an opportunity to |
| 2 |
come up with a good optimization doc but the paper is interesting. The |
| 3 |
answers might not be the same for P4 vs. AMD vs. sparc vs. Apple. |
| 4 |
Maybe a suite of files that get compiled, generate the numbers and |
| 5 |
instruct you what might work best? |
| 6 |
|
| 7 |
Interesting info. thanks. |
| 8 |
|
| 9 |
- Mark |
| 10 |
|
| 11 |
On 9/3/05, Volker Armin Hemmann <volker.armin.hemmann@××××××××××××.de> wrote: |
| 12 |
> On Sunday 04 September 2005 05:27, waltdnes@××××××××.org wrote: |
| 13 |
> |
| 14 |
> > |
| 15 |
> > That list shows pentium4, mmx, sse, and sse2. Also, if you have *ANY* |
| 16 |
> > version of sse available, you can improve performance by running floating |
| 17 |
> > point math via sse, rather than 387 instructions. I recommend... |
| 18 |
> > |
| 19 |
> > CFLAGS="-O2 -pipe -fomit-frame-pointer -march=pentium4 -mmmx -msse -msse2 |
| 20 |
> > -mfpmath=sse" |
| 21 |
> > |
| 22 |
> > |
| 23 |
> |
| 24 |
> emm. I would not do this. |
| 25 |
> |
| 26 |
> |
| 27 |
> -mfpmath=sse seems to be slower than -fpmath=387 |
| 28 |
> |
| 29 |
> http://www.anandtech.com/mac/showdoc.aspx?i=2436&p=5 |
| 30 |
> |
| 31 |
> has the numbers/made the experience. |
| 32 |
> |
| 33 |
> It seems, that gcc is not he best optimizer in the world ;) |
| 34 |
> -- |
| 35 |
> gentoo-user@g.o mailing list |
| 36 |
> |
| 37 |
> |
| 38 |
|
| 39 |
-- |
| 40 |
gentoo-user@g.o mailing list |
Archives