amd venice 3500+ and 3800+ preview

[saaya]

hmmmmmm

well they didnt run memtest or prime at all afaik, i guess we will have to find out about this ourselves :/

[saaya]

Specs:

DFI 939 Ultra

MSI X800XT 520/560

OCZ 3700 2.5-4-4-8 @ 300

All CPUs ran at 300x8 2400MHz

this is an average result!

there was no bad/buggy run

cant wait to see oppainters review

[Samael]

So there is a difference after all

The new instructions maybe boosted it a bit.

[saaya]

Αρχική απάντηση από Samael [Σήμερα, στις 05:53]

So there is a difference after all

The new instructions maybe boosted it a bit.

you think its the sse3 that gave it a boost? hmmm but the venice is also faster in 2k1 (only slightly, but nevertheless) and rdmark2001 def doesnt have sse3 support...

i think they speeded up something else... maybe they improved the crossbar

[Samael]

Now that i am thinking of it again....it might be from the so called "enhanced memory controller"

Did anyone run some unbuffered with the same settings?

[saaya]

Αρχική απάντηση από Samael [Χθες, στις 12:51]

Now that i am thinking of it again....it might be from the so called "enhanced memory controller"

Did anyone run some unbuffered with the same settings?

no, but sandra arythmetic is slightly higher than winchester and sandra multimedia is the same as clawhammer or winchester. i think memory bandwidth was higher with the venice, but dont know how much. pc ice will make some runs with an fx53 as well later i guess and compare it to his venice results.

dont know if he ran unbuffered... but i dont think so.

[saaya]

http://www.xtremesystems.org/modules.php?op=modload&name=Reviews&file=index&req=showcontent&id=2&page=2

final article is up! its a short preview with only 3dmark 2001 2003 and 2005 results since they had so little time to play with the cpus... :/

but very interesting! the venice beats the winchester in all tests and in one test the venice beats the clawhammer at the same clockspeed by almost 3%!!!

does anybody know a different cpu test that uses sse3?

[SLN]

encoding (divx xvid) tests should have SSE3 support (at least they say they do ), and should provide a real life useful test.

[saaya]

Αρχική απάντηση από SLN [Σήμερα, στις 16:50]

encoding (divx xvid) tests should have SSE3 support (at least they say they do ), and should provide a real life useful test.

hmmm but hard to test... i dont know how to test and benchmark with xvid, or does it have a benchmark included?

[SLN]

I don't know of any benches... Check the Anandtech/Toms reviews that include XviD, DivX benches.

To my knowledge, it's pretty straightforward, they just encode a certain portion or all of a DVD movie using autogordianknot (or gordian knot) from www.doom9.org. and use the fps to measure performance. Codecs have been advertising SSE3 support since the Prescott release. I don't know how much they take advantage of any optimization (or if they DO utilise them at all for that matter).

[saaya]

thx, i will check it out... if i get to review a venice or san diego i will definately use those tools as well!