Will 3DMark07 be a popular benchmark tool?

[Nick[FM]]

Nick, I think the problem some people have (at least, that's mine!) is that there is no combined SM2.0.+SM3.0. score combined on screen. As such, most people aren't going to give their "GPU" score even for GPU reviews or rumours.

So, if you guys only have DX10 benchmarks in 3DMark07, just make sure to have a final number for the GPU-only part. That should clearly be even more of an obvious design choice as with 3DMark06.


Uttar
P.S.: Neat features I wouldn't expect in 3DMark: GPU-based soundwave raytracing and occlusion?

So you mean that in addition to the 3DMark Score and CPU Score we should have a Graphics (or GPU) Score?

Currently we have no plans to have GPU-based soundwave raytracing and occlusion tests.

Cheers,

Nick

 

[trinibwoy]

So you mean that in addition to the 3DMark Score and CPU Score we should have a Graphics (or GPU) Score?

To re-iterate a point how about you guys get rid of the CPU score and just leverage multiple CPU cores in the 3D engine itself? You asserted earlier that the CPU score is there to represent the impact of multiple cores on future gaming performance. Like someone said earlier, why not just build an engine that benefits from multiple cores. It would be a much more valid test and would support Futuremark's stance on the issue.

If you guys aren't doing it why should we expect to see it from game devs anytime soon? And if we aren't gonna see it from game devs then there's no need for a CPU test in 3dmark since it doesn't convey particularly relevant information.

 

[Arun]

So you mean that in addition to the 3DMark Score and CPU Score we should have a Graphics (or GPU) Score?

Well, this should illustrate it better. Please excuse one of the worst photoshopped-with-paint pictures in history:

Uttar

 

[Nick[FM]]

Well, this should illustrate it better. Please excuse one of the worst photoshopped-with-paint pictures in history:

Uttar

Ok! The next 3DMark we will be DX10 only, so I am not so sure if we need to have separate SM-Scores.. But in any case, thanks for the good feedback (and artwork ).

Cheers,

Nick

 

[Nick[FM]]

To re-iterate a point how about you guys get rid of the CPU score and just leverage multiple CPU cores in the 3D engine itself? You asserted earlier that the CPU score is there to represent the impact of multiple cores on future gaming performance. Like someone said earlier, why not just build an engine that benefits from multiple cores. It would be a much more valid test and would support Futuremark's stance on the issue.

If you guys aren't doing it why should we expect to see it from game devs anytime soon? And if we aren't gonna see it from game devs then there's no need for a CPU test in 3dmark since it doesn't convey particularly relevant information.

If we would include the CPU workload in the graphics tests (or vice versa), we wouldn't be able to show pure GPU and pure CPU performance in gaming environments. We could of course make them all in one test, but currently we have no plans to take that route. Anyway, thanks for the good feedback.

Cheers,

Nick

 

[trinibwoy]

If we would include the CPU workload in the graphics tests (or vice versa), we wouldn't be able to show pure GPU and pure CPU performance in gaming environments.

Granted, but there is no such thing as pure CPU performance in a gaming environment - that's an oxymoron if I ever saw one

 

[Shtal]

A. A new 3DMark is hardly ever comparable to an old game, and I am not really sure what you mean with this question?

What I mean is how 3DMark07 will stress single core, dual core or multi-core in reality? when you run actual game tests but not CPU tests.
"It never happened in 3DMark03/05/06" Only in 3Dmark06 had CPU test but that doesn't count"

 

[Ateo]

As long as 3Dmark tells me nothing about how my games will run, I'm not getting it.
Last time I ran 3Dmark was on '03 and I learned it didn't tell me anything about gaming preformance.
So I don't see what this synthetic benchmark has to offer me...besides a E-penis number.

 

[Orbitech]

Well. A Quad Core System with a 8800GTX scores more than my Dual Core System with 8800GTX SLI. Maybe quad core will make that much difference in the future. But I doubt it.

Rofl... Are you sure? In that case I'll cancel my second 8800gtx purchase.. Quad core here I come..
Hilarious and right on target.
And to answer OP question.. Of course it will remain popular.. It has become a major marketing tool to be that easily neglected, which I suspect was Futuremark/Madonion's purpose from day 1..

 

[Shtal]

If we would include the CPU workload in the graphics tests (or vice versa), we wouldn't be able to show pure GPU and pure CPU performance in gaming environments. We could of course make them all in one test, but currently we have no plans to take that route. Anyway, thanks for the good feedback.

Cheers,

Nick

How about something like you did with 3DMark2001/SE
Game test 1 Low+High Detail was Car test.
Game test 2 Low+High Detail was Dragon flying test.
Game test 3 Low+High Detail was Max Payne gun test.


For 3DMark07 something like this!
Have game test 1 scene 1 - Pure Graphic test: "as usual" then,
Have game test 1 scene 2 - Mix CPU+GPU depended test.
----------------------------------------------------------------------------------------
same goes for:
Have game test 2 scene 1 - Pure Graphic test:
Have game test 2 scene 2 - Mix CPU+GPU depended test.
----------------------------------------------------------------------------------------
same goes for:
Have game test 3 scene 1 - Pure Graphic test:
Have game test 3 scene 2 - Mix CPU+GPU depended test.
----------------------------------------------------------------------------------------
This way 3DMark07 will have opportunity to stress Video Processing Unit (GPU)
Plus - Central Processing Unit(CPU): And then give overall score for both of them as one score.

best regards,

Shtal....

[Edit: You could always add "1 or 2" CPU test(s) after game test 3 scene 2 - Mix CPU+GPU depended test.]

 

[pjbliverpool]

3DMark07 to be DX10/Vista only:

http://www.yougamers.com/blogs/nick_r/

Im quite pleased about this but im guessing many won't be.

 

[Kaotik]

3DMark07 to be DX10/Vista only:

http://www.yougamers.com/blogs/nick_r/

Im quite pleased about this but im guessing many won't be.

Considering Nick said it in this very thread in January, it's hardly news

 

[fallguy]

Maybe it will actually be useful this time. I doubt it though.

 

[Sobek]

How about something like you did with 3DMark2001/SE
Game test 1 Low+High Detail was Car test.
Game test 2 Low+High Detail was Dragon flying test.
Game test 3 Low+High Detail was Max Payne gun test.

See this is what I liked about '01, the individual tests seemed to stress individual system parts seperately. It was weird, but at the same time extremely usefull.

Game test 1 showed the largest changes when overclocking your videocard.
Game test 2 showed the largest gains from overclocking your CPU.
Game test 3 seemed to be a strange hybrid test, where overclocking one or the other didn't seem to have any massive impact, but in tandem a good overclock on both your CPU/GPU seemed to show massive gains.

I know, sounds stupid. But I lived by that process back in the day and it never failed to perform. I've always wanted a true successor to 3dmark01, not this rounded, shiny, bubbly glowy over-bloomed 6 hour long test cycle we've had so far (no offence to you FM guys, it's amazing nonetheless, but I don't think it's a benchmark in the strictest sense. Especially not with the ridiculous way the CPU can affect the overall score).

Fingers crossed for '07! :smile:

 

[MrBelmontvedere]

IMO it's good that they are stressing dual/quad core performance. while this is misleading for current single-threaded games, in the future I think more games (especially those that put out high end gfx) will be multi-threaded. and to me, that's what this kind of benchmark should be about: showing how a given system might perform in FUTURE games. after all, if you want to see how your system performs in current games, just run FRAPS and see for yourself in the actual game.

 

[Shtal]

There's no reason to go backwards, if you know what I mean.. http://www.yougamers.com/blogs/nick_r/

lol: So this benchmark will be build / help to determine who is better R600 vs. G80
Will be quite interesting if this benchmark be favor to ATI or Nvidia (Example like in 3DMark06; Nividia wins in SM2 and ATI wins in SM3.)

 

[Rangers]

Will 3DMark07 be a popular benchmark tool?

Hmm, I'm gonna go ahead and guess yes

LOL.

 

[Frank]

Hmm, I'm gonna go ahead and guess yes

LOL.

Agreed. It is the single benchmark for the masses.

 

[Reverend]

Yes, many websites will use 3DM07 since this is a thread about "popularity".

Wrong forum IMHO btw.

 

[Sunday]

What's so special about the number 3? Some of your deductions are really strange

And why in God's name would another DX9 test be of any value at all? The DX9 card results would be worthless since they could not run the other two tests at all.

exactly! If they make any game as DX9 test, they've buried 3DM07!