fastest floating point operation as possible

On Mon, 09 Feb 2004 02:00:14 GMT, Greg Berchin
wrote:

On Sun, 08 Feb 2004 15:58:03 +0100, in wrote:


Just this morning I rearranged everything so that all "div"
operations were isolated. I got an improvement of just 0.8%. Of
course, this is with standard 387 instructions.

Yes, but all that concerned the P4, Intels compiler and SSE2.

I have a friend with a 1.7 GHz P4 machine, don't know his memory
speed. My XP2100+ machine runs the same application 19% faster
than his (2814 seconds vs. 2792, for the short test that we
tried). Remember that my system doesn't have DDR memory.

I'd say that is pretty good for the P4, suggesting your code isn't in
as bad shape as you fear. Try Intels compiler.

Aw, crud. I know at least a LITTLE bit about computers, and I got
fooled. What do people who know nothing do?

They buy high clockrates, wimpy ram, wimpy hd, VERY wimpy graphics and
low performance. Usually from Dell.

ancra

On Mon, 09 Feb 2004 16:44:50 +0100, in wrote:

I'd say that is pretty good for the P4, suggesting your code isn't in
as bad shape as you fear. Try Intels compiler.

Today I replaced all of my code with the 3DNow!/MMX equivalents
from the AMD sample code included with Processor Pack 5. My code
now runs in 1/14 the time that it used to, or about 58% of real
time! (That is, it now takes 0.58 seconds to process 1 second's
worth of data, whereas it used to take more than 8 seconds to
process 1 second's worth.)

Wow.

Greg

On Tue, 10 Feb 2004 02:01:49 GMT, Greg Berchin
wrote:


Today I replaced all of my code with the 3DNow!/MMX equivalents
from the AMD sample code included with Processor Pack 5. My code
now runs in 1/14 the time that it used to, or about 58% of real
time! (That is, it now takes 0.58 seconds to process 1 second's
worth of data, whereas it used to take more than 8 seconds to
process 1 second's worth.)

Wow.

Terrific!

ancra