• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

ComputeMark v1.2 - CrossFire/SLI support, bench

Soldato
Joined
7 May 2006
Posts
12,183
Location
London, Ealing
computemarkbench.jpg

http://www.czechgamer.com/download/3617/ComputeMark-v12-DX11-ComputeShader-benchmark.html
 
GTX4xx are easier to get performance out of them. Is this benchmark based on open CL? Or is it optimised for both?

Anyway, it doesnt show any kind of realistic workload, its really hard for the 5xxx series to get near theoretical performance. Fermi may be disappointing, but its GPGPU side is stronger than ATi for now.
 
That 5970 looks extremely powerful compared to the 480, wonder what one of those new OC'd 4gb Sapphires would get!

The question is, will there be any apps using Direct Compute any time soon?
 
Can you post any links that actually show this, rather than stating opinion as fact?

It can't even be opinion when the numbers in front of him say otherwise.

It's like saying "in my opinion, that number 3 on the wall doesn't actually say "3" it ACTUALLY says 10, because that's my opinion".
 
Its based on stuff that i read on tech forums.

2 minute google yielded this, http://www.anandtech.com/show/2977/...tx-470-6-months-late-was-it-worth-the-wait-/6

The Open CL benchmarks show the GTX at a great advantage, for now.

It depends greatly on how the code is written, if it's generic code the GTX 480 may be faster, but if code is optimized for each platform that may not be the case.

e.g if you look at this thread they gave the 5870 a 50% speed increase through some small code changes and eventually had the 5870 running faster than the 480 GTX where's before it was much, much slower.

http://forum.beyond3d.com/showthread.php?t=57042
 
It depends greatly on how the code is written, if it's generic code the GTX 480 may be faster, but if code is optimized for each platform that may not be the case.

e.g if you look at this thread they gave the 5870 a 50% speed increase through some small code changes and eventually had the 5870 running faster than the 480 GTX where's before it was much, much slower.

http://forum.beyond3d.com/showthread.php?t=57042

Exactly. The question is whether OpenCL is completely unbiased as to hardware and to whether code can be optimised for a wide variety of applications without too much trouble/money.

It also poses the question on whether real-world apps are limited in the way that they can be optimised.

We know that nvidias developer relations are excellent, lets hope that AMD rises up to showc ase the potential to their hardware.

Anyway, time will tell, it should be interesting to watch.
 
Last edited:
TB's Burn results do look a little odd?
My results are about half what Userunknown has, which considering he has 2x 5870, is not bad :)

Results in sig....
 
Back
Top Bottom