Interesting, I find the situation a little ironic

[Joe DeFuria]

The fact of the matter is it will be used improperly, regardless of how many times you post that it shouldn't be. OEMs will trumpet these numbers on the back of their boxes, and lay people in Best Buy will be "mislead" into thinking product A is 2x faster/better, when it isn't in real life, not now, and not in the reasonably forseeable future.

Yes, and that is true of any benchmark on the planet, including "so called" game benchmarks. Such as UT2K3 Fly-bys. So why are we singling out 3D Mark?

Measuring how "GOOD" a product is has a subjective element to it. Putting "DX9 Compliant" on a box for the lay people in Best Buy doesn't give them a clue as to how much "better" that makes the card than a DX8 card either.

The 3DMark score attempts Quantify that.

Is a Radeon 9700 3X as good as a GeForce4 Ti? I bet some people would say yes, and some would say no, and some would say about right. It depends on what they do with the card, what games and apps they run, how long they plan to keep it, etc.

 

[Himself]

IMO, the only cards that should be tested on 3dmark03 are the GFX and 9700 generations and beyond. Do we really care if some DX8 card gets 2fps or 4fps? Seems like much ado about nothing.

 

[Bjorn]

IMO, the only cards that should be tested on 3dmark03 are the GFX and 9700 generations and beyond. Do we really care if some DX8 card gets 2fps or 4fps? Seems like much ado about nothing.

You're right. And that's the problem. I think it's a bit to early to have 2 out of the 4 game tests giving out 1-4 fps. This on high end DX8 hardware. You shouldn't need a DX9 card to get playable framrates in the DX8 tests (dunno if you should use playable when talking about a pure benchmark though ).

Why keep the DX8 tests at all ? Better to use DX9 in all but the first test then.

 

[Joe DeFuria]

You shouldn't need a DX9 card to get playable framrates in the DX8 tests (dunno if you should use playable when talking about a pure benchmark though ).

That's just it....we are talking about a benchmark purposely designed to stress the GPU. This has nothing to do with playability.

Now, maybe you believe that the test shouldn't stress the GPU so disproportionally?

If not, you have to keep in mind that this benchmark needs to last probably 18 months to 2 years. A card that was released 6 months ago is "already" getting 30 FPS or so in those tests. What are cards 2 years from now going to be pulling? We we be complaining then that the test isn't stressful enough?

It's a conscious decision FutureMark had to make. While I would agree that I think ONE of the DX8 tests should have been "less stressful", I would prefer both tests to be stressful, than having both running at 30-60FPS on today's DX8 cards.

 

[Himself]

That's not really futuremark's problem, I would say the problem is the lack of a intensive DX8 benchmark to show what games in the near future are going to be like. (or so goes the theory, I have to be convinced that games will ever get much past DX7 until everybody, and I mean everybody, has a DX8 card. )

There is a gap where a 3dmark2002 should have gone in terms of market lag. 3dmark2001 was mostly DX7 with a taste of DX8, 3dmark03 is mostly DX8.1 with a taste of DX9. Seems consistant enough to me. It's not like 3dmark2001 has dropped off the place of the planet, it's still there for older cards.

 

[Bjorn]

That's just it....we are talking about a benchmark purposely designed to stress the GPU. This has nothing to do with playability.

Now, maybe you believe that the test shouldn't stress the GPU so disproportionally?

If not, you have to keep in mind that this benchmark needs to last probably 18 months to 2 years. A card that was released 6 months ago is "already" getting 30 FPS or so in those tests. What are cards 2 years from now going to be pulling? We we be complaining then that the test isn't stressful enough?

It's a conscious decision FutureMark had to make. While I would agree that I think ONE of the DX8 tests should have been "less stressful", I would prefer both tests to be stressful, than having both running at 30-60FPS on today's DX8 cards.

I basically agree with you. I'm not expecting to get 50 fps at 1024 in Doom3. Far from it. But i don't expect 1-2 fps either. Problem is, would rather have had two DX9 tests then instead of one DX8 test that gives you 1-4 fps. GT4 and another one that was more like the nature test was when 3D Mark 2001 was released. Something that brought the R9700 and GF FX to it's knees. (5-10 fps or something like that).

If not, you have to keep in mind that this benchmark needs to last probably 18 months to 2 years. A card that was released 6 months ago is "already" getting 30 FPS or so in those tests. What are cards 2 years from now going to be pulling? We we be complaining then that the test isn't stressful enough?

As i have said before, the GFX is actually faster in the DX9 test then in the GT4. And also over 30 fps. So i'm already complaining that that test isn't stressful enough