Will 3DMark07 be a popular benchmark tool?

[Shtal]

We have heard that Futuremark's 3D Mark 07 will be fully Quad-Core Optimized, even in the Game Tests. Users will gain more frame rates with a Quad Core Processor even in graphics tests. Current beta version seem to indicate that there are 3 Games Tests. http://www.vr-zone.com/?i=4358
__________________________________________________________________

I remember the days when MadOnion/FutureMark introduce 3DMark2001 based on DX8.0 API for future DirectX8.0 title games. It was solid good real popular benchmark that proved and - even used for some future game title releases. Since then later with 3DMark03/05/06 - in my opinion was pure synthetic based faulty not as good as 3DMark2001SE.

My question is will 3DMark07 utilize all 4 cores (Quad Cores), - that can show real/major difference's vs. 2 cores (dual core) when running some tests; and also by how much 3DMark07 can reflect for future game titles in real life.

I would like to add will you be able still run DX9 hardware (I heard it is only going to be DX10) people say we have already 3 versions of DX9-3DMark.

 

[Mendel]

I would like to add will you be able still run DX9 hardware (I heard it is only going to be DX10) people say we have already 3 versions of DX9-3DMark.

Seeing as 07 is a vista only benchmark, I wouldn't be surprised if there was at least one dx10 only test.

 

[trinibwoy]

Why don't you post these questions at Futuremark's forums? They are the only ones who can answer them after all.

 

[Sobek]

I remember the days when MadOnion/FutureMark introduce 3DMark2001 based on DX8.0 API for future DirectX8.0 title games. It was solid good real popular benchmark that proved and - even used for some future game title releases.

Someone else said pretty much the exact same thing in another 3dmark07 thread just yesterday I believe.

The engine 3dMark01 was based on was never used in any game, and as someone previously pointed out, was pretty damn different to the version fo MaxFX used in the original Max Payne...

 

[Dave B(TotalVR)]

FYI

http://www.3dchips-fr.com/actualites/11923/3d_mark_2007_des_images_maj.html

 

[Kaotik]

FYI

http://www.3dchips-fr.com/actualites/11923/3d_mark_2007_des_images_maj.html

I believe those are shots from a demo/bench FM made for Intel, not from 3DMark07

 

[Demirug]

Someone else said pretty much the exact same thing in another 3dmark07 thread just yesterday I believe.

The engine 3dMark01 was based on was never used in any game, and as someone previously pointed out, was pretty damn different to the version fo MaxFX used in the original Max Payne...

But at least it used a game engine. Today FutureMark tests are more written like graphics demos.

 

[kyetech]

Why don't you post these questions at Futuremark's forums? They are the only ones who can answer them after all.

Yes or:

At the very least, why didnt he post his question in the thread that has already been started in this forum just below this one?

seems silly to have two threads on something that no-one knows anything about!

 

[Shtal]

Someone else said pretty much the exact same thing in another 3dmark07 thread just yesterday I believe.

The engine 3dMark01 was based on was never used in any game, and as someone previously pointed out, was pretty damn different to the version fo MaxFX used in the original Max Payne...

To be honest truth I did not knew someone else already asked that question!
And I did not paid attention if this type tread already exist!
Sorry!

 

[rwolf]

3DMark is a great product and has been a great tool to benchmark video cards. With DX10 arround the corner it will become even more important as there won't be any games for a while.

 

[Ateo]

3DMark is a great product and has been a great tool to benchmark video cards. With DX10 arround the corner it will become even more important as there won't be any games for a while.

Only if you consider comparing 3Dmark scores to each other.
It tells nothing about real world in-game preformance at all.

Just look at how 3Dmark06 and the CPU score is misleading.
No games n the market get such benefits from having dual/quad-core CPU's.
And when(if?) they do 3Dmark06 won't be the 3Dmark being used to test preformance.

But I guess it's good for E-penis messuarment...

 

[Frank]

Q: how many games will take advance of 4 CPU cores in the near future?

A: Perhaps a few, in a small way. What does it matter? There isn't that much they want to do with them in any case, as most of the consumers will only have a single CPU core anyway and it isn't cost-effective to spend the time and money on it. Scrap multiplayer as a start.

 

[Shtal]

The only version I like of 3DMark was 2001SE. It had Game1 low detail test which take advantage of faster CPU; then Game1 High detail test which benefit from faster video card. - Same goes for Game 2 test and Game 3 test.

Since 3DMark07 also has 3 Game tests, it might have same approach as 2001SE.

The only thing it would be nice in 07 if 1 out of 3 tests had DX9-SM3.0 and other 2 would had DX10-SM4

 

[trinibwoy]

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.

 

[Shtal]

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.

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.
Game test 4 was Nature Test

And for DX9-SM3 to include is because in my opinion 3DMark06 DX9 did not reflect any real world / reality of actual games.

 

[Davros]

"I remember the days when MadOnion/FutureMark introduce 3DMark2001 based on DX8.0 API for future DirectX8.0 title games."

nyone remember their first benchmark - final reality ?

 

[no-X]

I don't think that 3DM01 was a good benchmarking tool. It was popular, but not very objective. GeForce = good score / Kyro II, Radeon, Voodoo 5 = low score (e.g. GF256 gives much better results than other mentioned boards, but majority of games runs much smoother on Radeon or Kyro II...). Anyway, 2001 was far more objective than 2000, which used strange combination of 16bit rendering and HW TnL :smile:

 

[trinibwoy]

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.
Game test 4 was Nature Test

So what?

And for DX9-SM3 to include is because in my opinion 3DMark06 DX9 did not reflect any real world / reality of actual games.

What makes you think 3dmark07 will be any different?

 

[Shtal]

So what?

At least I will not hear negative responds from you (I don't like being wrong)

Example you said "Some of your deductions are really strange"

What makes you think 3dmark07 will be any different?

I don't! Just my own thoughts

 

[Shtal]

Q: how many games will take advance of 4 CPU cores in the near future?

A: Perhaps a few, in a small way. What does it matter? There isn't that much they want to do with them in any case, as most of the consumers will only have a single CPU core anyway and it isn't cost-effective to spend the time and money on it. Scrap multiplayer as a start.

You may be right! But not exactly I would agree with you (Take a look at the results with multi-core)

[What if FutureMark found a way to split work load 4-way tread to do its own calculation dependency (Example old Comanche 4 helicopter's single treaded CPU dependency game) I know its not 4-treaded "bad example" ]

Typically before entering a game, most gamers will close all their open apps before launching the game. The reason is simple: more performance. But what if you could keep all your apps open, and even do a little bit of encoding while you’re getting your game on? This is what we set out to do with our multi-tasked gaming tests. With them we ran our standard suite of MP3 and WME 9 tests while also running our gaming benchmarks.
As you can see in Oblivion, performance improved dramatically under this scenario in Oblivion. At 800x600 the quad-core CPUs delivered over double the performance of the Core 2 Extreme X6800, nearly matching the performance numbers we saw a few pages ago in Oblivion (80 fps in this case versus 82.5 previously in outdoors, and 47 versus 51 in city). At 16x12 the burden shifts to the graphics card and the performance hit is gone completely http://www.firingsquad.com/hardware/intel_core_2_extreme_qx6700/page14.asp