AMD optimized Flask is here!!

[Alfred]

AMDZone

 

[Sephiroth_IX]

Good. No more P4 domination.

 

[Soccerman]

actually seph, this one still gets beaten by Intel's version of Flask (with SSE2 optimizations). I think the explaination for this is that SSE 2 is more powerful, and the Intel P4 is alot more optimized around SSE2.

as we all know as well, the K6-2 benefits far better persentage wise with 3DNow!, then the Athlon. this is because the Athlon was not made to benefit alot through 3DNow!

3DNow! was originally made to benefit the K6-2.

and also, the P4 probably benefits with a nearly double helping of FSB speed and RAM bandwidth.

Tom's hardware has the dirt on this one btw..

 

[Alfred]

we have to wait for the version that directly comes from AMD which Tom has mentioned.

 

[Soccerman]

he also mentioned that the one he had was pretty much the bulk of the benefit that you'll see from optimizing for AMD processors.

 

[office boy]

LAme.
This http://www.dvd-digest.net/downloads/files/encode/flaskpx3_004e_final.zip version is faster.

 

[A5]

I thought you guys were talking about a drinking flask...now that would be funny.

 

[GL]

Am I missing something here? Looks like the AMD ties the P4. Tom's bar graph shows the low quality benchmarks of the P4 beating the AMD benchmarks which are all in high quality mode. In high quality mode, the difference is about 2 fps between an Athlon at 1.47GHZ and a P4 at 1.5GHz.

EDIT
Oops thought the Athlon was at 1.4GHz...guess it isn't really a tie but just really close. But apparently there is still more 3dnow! code that might tweak a frame or two a second out of the code rendering the benchmark a toss up between the P4 and Athlon.

-GL

 

[Tonec]

52

 

[Soccerman]

it's weird.. Tom included in the P4 benches with and without the optimizations at high quality, AND also included the benches for low quality only with the SSE2 optimizations.

it makes that CPU look alot better if you ask me (at first I thought the comparison said that the P4 hit 22.85 and the Athlon hit 16.97).

now that I reread this bench, the only valid benches for comparison that include optimisations are the ones where the P4 hits 18.96 and the one where the Athlon hits 16.97.

this leads me to believe STILL that 3DNow! has less of an impact on the Athlon as it did on the K6-2, and also, that AMD's Athlon is limited by fsb and memory bandwidth (to a lesser degree then I originally thogh for this benchmark, however further into the review, you see fsb and memory speed playing a LARGE role, which is why the P4 does well in those benches).

 

[nickburns]

i am with A5 on this one, i expected to see an AMD flask that would hold some booze, i am dissapointed