ShaderMark NV40 vs R420

[Bob3D]

http://www.nordichardware.com/reviews/graphiccard/2004/r420/index.php?ez=13


When Microsoft releases DirectX 9.0c we will be able to evaluate Pixel Shader 2.0a, 2.0b and 3.0.

The results speak for themselves. NV40 isn't capable of running seven out of the 23 tests. In those tests where both cards are able to run the X800 XT is the faster card in all test but one. Eventually this might change when DirectX 9.0c is released and NV40 can use Pixel Shader 3.0 in the tests. With all probability NV40 will also be able to run all tests when the next update of DirectX will be released.


So, DX9.0c can give NV40 a performance boost or PS 3.0 can make the performance even worse?

 

[Tim Murray]

The answer you definitely did not want to hear: Both. If a dev uses PS3.0 correctly in a situation where it makes sense to use branching/swizzling/etc., sure, he can improve performance no problem. If he's stupid and uses PS3.0 for the sake of using PS3.0, hello performance armageddon!

 

[trinibwoy]

Ouch. How is it possible for NV40 to be so completely humiliated in those synthetic tests yet have a much better showing in actual games? And why can't the NV40 run many of the tests given it's full SM2.0 compliance?

 

[digitalwanderer]

What about when dx9.1 comes out? Won't nVidia wipe the floor with ATi then? :|

 

[Joe DeFuria]

What about when dx9.1 comes out? Won't nVidia wipe the floor with ATi then? :|

Yeah....but that won't be until at least July!!

 

[T2k]

What about when dx9.1 comes out? Won't nVidia wipe the floor with ATi then? :|

9.1? There's no such a thing as MS cleared months ago, IIRC

Edit: wrong smiley

 

[Tim Murray]

Re: Previous two posts (before that bastard who replied while I was posting, YAAAAR!)

Best. Posts. Ever.

Anyway, in a vain atetmpt to keep this on topic, does the NV40 actually have all of the SM2.0 caps exposed in PS2.0, or do a lot of them (MRT, etc.) rely on PS3.0 support? That'd be an interesting way to force adoption...

 

[Tridam]

Ouch. How is it possible for NV40 to be so completely humiliated in those synthetic tests yet have a much better showing in actual games? And why can't the NV40 run many of the tests given it's full SM2.0 compliance?

I presume that this problem comes from a lack of some render target formats support. For example, NVIDIA doesn't support D3DFMT_R16F. Some tech demos are using this RT format as ATI has support for it for a while. However NVIDIA supports D3DFMT_R32F and D3DFMT_G16R16F. If the problem is the lack of D3DFMT_R16F support it's easy to use D3DFMT_R32F or D3DFMT_G16R16F instead. If this is the problem then I presume that the next revision of Shadermark will use one of these RT formats for NVIDIA GPU.

 

[Bob3D]

That's sad
I really don't want to go with ATI cards, but looking at these benchs...
I can't understand why NV40 with all these "sweet shaders" can't keep the performance at level with ATI new cards. In all shaders benchs we see nvidia losing.
Looks like this will be the same NV30 vs R300 shader scenario

 

[trinibwoy]

Personally I'd like to see benches where NV40 can run all the tests to get a more complete picture. It was a lot easier to understand NV30s shortcomings but this time around I can't seem to grasp why the X800XT is handing the NV40 its ass in a bucket in all these tests. Where is the architectural flaw?

 

[Ante P]

Guys, I just ran those benchies with the new 61.11 drivers and scores increased a bit in some of the tests.

In FillrateTester the results went through the roof and increased by several thousands of MP/s when it comes to the FFP tests.

 

[Tridam]

BTW there are some errors in the nordichardware fillrate table. The 6800 texturing results and some others are wrong.

 

[Ante P]

Anyway, in a vain atetmpt to keep this on topic, does the NV40 actually have all of the SM2.0 caps exposed in PS2.0, or do a lot of them (MRT, etc.) rely on PS3.0 support? That'd be an interesting way to force adoption...

I'm thinking no. Looking at DXcaps it seemed like some of the stuff that could have been exposed under 2.0 (2_x) isn't right now while some other new stuff (like MRTs) are.
Can't remember what though can check again later but I'm gonna be away for some time now.

 

[Ante P]

BTW there are some errors in the nordichardware fillrate table. The 6800 texturing results and some others are wrong.

See above.

 

[Diplo]

I really don't want to go with ATI cards, but looking at these benchs...

You should go with whatever is best at the time.

I can't understand why NV40 with all these "sweet shaders" can't keep the performance at level with ATI new cards.

I have a feeling those tests aren't quite right, but need to do some checking...

 

[Tridam]

BTW there are some errors in the nordichardware fillrate table. The 6800 texturing results and some others are wrong.

See above.

I'm not fast enough

I'm sure that the problem has nothing to do with the new drivers. With 60.72 I have the same numbers than with 61.11.

 

[Ante P]

I have a feeling those tests aren't quite right, but need to do some checking...

They are wrong, I've updated them in an upcoming article:

By the way the second 6800 Ultra (the one on the right) is the Extreme version.

 

[digitalwanderer]

What about when dx9.1 comes out? Won't nVidia wipe the floor with ATi then? :|

Yeah....but that won't be until at least July!!

Touche' Dr. DeFurious, your wit is sharp and quick!

You just tied Baron for Pepsi out me nose laugh and beat me at me own joke, a definate two pointer!

 

[Bob3D]

really
http://www.anandtech.com/video/showdoc.html?i=2044&p=9

 

[Bob3D]

can anyone here write some PS 3.0 shaders to run with NV40?