NV40 3DMark 2003 scores revealed -theinquirer

[PaulS]

The odds of them getting the synthetic benches right are very low, considering that the only real public information is the final score and that the synthetic benches don't contribute to that. As it stands, though, I think that's a fake.

Dumb question, but if we don't know what the synth scores are and if they WOULD have faked them how would we have the slightest idea if they were faked or not? (I guess I'm trying to figure out if there is a way to determine what the rough synths scores should be, I don't know of any.)

I'd imagine that would be obvious, Dig.

 

[digitalwanderer]

I'd imagine that would be obvious, Dig.

I'm not seeing the forest for the bloody trees again then!

Which is obvious; what the synths would be or that we just don't know?

Dan-510fps in 320x240? I don't think 400 sounds very realistic at 1024x768 then, but.... :|

I'm curious how many websites will use 3dmark03 in their nv40 previews

The Dig spits at the ground in disgust, then remembers who's forum he's at and quickly bends down and wipes it up with a tissue and settles on looking a bit annoyed-with-the-state-o-the-industry instead.

All of them, bloody all of them. The ones who were against it will use it 'cause it will make the nv40 look better, the ones who supported it will use it since they support it, everyone else will be bloody analyzing the results seventy-eleven ways to Sunday to see who's cheating at what or why it isn't reflective of something or other.

 

[dan2097]

Dan-510fps in 320x240? I don't think 400 sounds very realistic at 1024x768 then, but....

Well it shows that if the geforce 5800 ultra is infinitly fast it should get that score in GT1 when paired with a highly overclocked cpu

 

[Rys]

60.72's aren't FutureMark approved. Won't we be breaking the terms of the software by publishing results? 52.16 doesn't work with the card.

I don't think I will be. The 12500 figure is public knowledge and it's correct, no point repeating it and peeing FM off

Might use the PS and VS numbers though

Rys

 

[PatrickL]

That just means FM needs to put out a new patch.

 

[dan2097]

That just means FM needs to put out a new patch.

Which they'll likely be doing anyway to add ps/vs 3.0 support

 

[jvd]

That just means FM needs to put out a new patch.

Which they'll likely be doing anyway to add ps/vs 3.0 support

why would they add ps/vs 3.0 support . Only 1 card maker supports it . If anything they'd have to put it in a non scored test . After all thats what they did with p.s 1.4

 

[PatrickL]

New patch is needed because it is actually easy to post Fm results with unapproved drivers as they have the perfect excuse with the fact that there is actually zero approved drivers supporting Nv40.

 

[Geeforcer]

That just means FM needs to put out a new patch.

Which they'll likely be doing anyway to add ps/vs 3.0 support

why would they add ps/vs 3.0 support . Only 1 card maker supports it . If anything they'd have to put it in a non scored test . After all thats what they did with p.s 1.4

Err, because more then one will support VS3.0 right away and eventually all will support PS3.0, much unlike PS1.4? Using your "logic", they should not have put PS2.0 support into 3dmark03 untill Nv30 came out, since prior "only 1 card maker support(ed) that."

 

[Mark]

That just means FM needs to put out a new patch.

Which they'll likely be doing anyway to add ps/vs 3.0 support

why would they add ps/vs 3.0 support . Only 1 card maker supports it . If anything they'd have to put it in a non scored test . After all thats what they did with p.s 1.4

Err, because more then one will support VS3.0 right away and eventually all will support PS3.0, much unlike PS1.4? Using your "logic", they should not have put PS2.0 support into 3dmark03 untill Nv30 came out, since prior "only 1 card maker support(ed) that."

IIRC, that's exactly why 3DMark03 was delayed for so long. Futuremark didn't see a need to get it out because "only 1 card supported it" (even though ATI had several cards which supported PS2.0... granted they all used the same R300 core).

 

[Dave Baumann]

A.) I'm not sure sure 60.72's won't be FM approved.

B.) Highly unlikely that any patches for '03 would add anything extra. There certainly wouldn't be anything added/changed to the game tests.

 

[digitalwanderer]

A.) I'm not sure sure 60.72's won't be FM approved.

Two things:

1. Are the 60.72s the same as the 60.32s and I just screwed the number up?

2. Has anyone tested the 60.7/32s on the FX to see if optimizations are still present or if the score is the same?

 

[Dave Baumann]

Sorry...

I'm not sure 60.72's won't be approved for potential products based of a chip designation greater than NV36....

 

[digitalwanderer]

Sorry...

I'm not sure 60.72's won't be approved for potential products based of a chip designation greater than NV36....

Thanks, even I can read what you ain't writing on that one!

(BTW-Damn it! )

 

[jvd]

Err, because more then one will support VS3.0 right away and eventually all will support PS3.0, much unlike PS1.4? Using your "logic", they should not have put PS2.0 support into 3dmark03 untill Nv30 came out, since prior "only 1 card maker support(ed) that."

using your "logic" one supported p.s 1.4 right away and eventually all supported p.s1.4 . The r200 , r300 , nv30 , r420 and nv40 all support p.s 1.4

Yet it didn't count in the scores .

Of course only one card supported p.s1 .0 so why should it be diffrent this time around .

Doesn't really matter to me . I don't trust a 3dmark score anymore than I trust an enron ceo with my money .

 

[Tim Murray]

Sorry...

I'm not sure 60.72's won't be approved for potential products based of a chip designation greater than NV36....

Oh, so they'll be approved for NV38 then?

 

[Dave Baumann]

Ahhhhh, crap, forgot about that one.

 

[I.S.T.]

ROFL!


Which one is the NV38 again? The 5950? Edit: I checked the main site and found out I was correct.

 

[Geeforcer]

Err, because more then one will support VS3.0 right away and eventually all will support PS3.0, much unlike PS1.4? Using your "logic", they should not have put PS2.0 support into 3dmark03 untill Nv30 came out, since prior "only 1 card maker support(ed) that."

using your "logic" one supported p.s 1.4 right away and eventually all supported p.s1.4 . The r200 , r300 , nv30 , r420 and nv40 all support p.s 1.4

Yet it didn't count in the scores .

Of course only one card supported p.s1 .0 so why should it be diffrent this time around .

Doesn't really matter to me . I don't trust a 3dmark score anymore than I trust an enron ceo with my money .

Only one chip from one manufacturer (R200) supported PS1.4 during 3dmark01 lifetime. Certainly, if multiple cards support VS3.0 and within a couple of months multiple IHVs support PS3.0 (IMG?) would be quite different from PS1.4 support. That said, I would agree that adding a new scoring test to mark03 would contradict FM justification for not making PS1.4 a scoring test: it would fracture the community and make the scores impossible to compare.

 

[KimB]

Only one chip from one manufacturer (R200) supported PS1.4 during 3dmark01 lifetime. Certainly, if multiple cards support VS3.0 and within a couple of months multiple IHVs support PS3.0 (IMG?) would be quite different from PS1.4 support. That said, I would agree that adding a new scoring test to mark03 would contradict FM justification for not making PS1.4 a scoring test: it would fracture the community and make the scores impossible to compare.

Except now any PS 3.0 algorithm should be able to be emulated in PS 2.x. This is what 3DMark03 does: it has multiple shader fallbacks for previous video cards. Why not add some PS 3.0 tests, and continue to have fallbacks for previous PS versions?