-=D700=- Post Heaven score here please!

[ChuckBlack]

So we've got scores for the D300 and D500 cards for the Heaven Benchmark, extreme preset. We're still short on the D700's score. Could someone with D700 GPU's run this benchmark. Yes, the CPU's have been different, this is by no means a scientific test, just some nerding-out. THANKS! A link to the benchmark is below...

http://unigine.com/products/heaven/download/

 

[Quash]

Indeed, would be nice if someone here know someone with a review unit, like anand

I expect them to score around 30FPS mark

 

[Coyote2006]

I wonder. These GPUs are said to be optimized for OpenGL but score that low?

 

[Anim]

why does 12 core (4 extra cores) with D500 (instead of D300 @ 8 cores) score less?

That makes little sense doesn't it? Assuming both Turbo Boost during the test.

 

[mrsavage1]

yeah this was expected. We were all speculating that the d500 were gonna perform worse than the d300 in games.

 

[Anim]

yeah this was expected. We were all speculating that the d500 were gonna perform worse than the d300 in games.

Still makes no sense, unless its the CPU in the 12 core.

D500 Specs
========
1525 Stream Processors (texture mapping units/unified shaders)
384-bit memory bus width
240 GB/s memory bandwidth
3 GB Vram GDDR5

D300 Specs
========
1280 Stream Processors (texture mapping units/unified shaders)
256-bit memory bus width
160 GB/s memory bandwidth
2 GB Vram GDDR5

 

[N19h7m4r3]

I wonder. These GPUs are said to be optimized for OpenGL but score that low?

They're for OpenCL, there's a difference. AMD is leagues ahead of NVIDIA in OpenCL, but NVIDIA is better at OpenGL.

AMD's OpenGL drivers have always been playing catch up, so we can only hope Apple is going to get them to sort that out.

 

[Coyote2006]

They're for OpenCL, there's a difference. AMD is leagues ahead of NVIDIA in OpenCL, but NVIDIA is better at OpenGL.

AMD's OpenGL drivers have always been playing catch up, so we can only hope Apple is going to get them to sort that out.

Thanks for clarifying this!

I've just done some OpenCL tests with Luxmark 2.1 and my GTX680 scores only 755.

But on OpenGL my Zotac is much faster thann the D300 or D500: FPS 53.4

 

[mrsavage1]

Still makes no sense, unless its the CPU in the 12 core.

D500 Specs
========
1525 Stream Processors (texture mapping units/unified shaders)
384-bit memory bus width
240 GB/s memory bandwidth
3 GB Vram GDDR5

D300 Specs
========
1280 Stream Processors (texture mapping units/unified shaders)
256-bit memory bus width
160 GB/s memory bandwidth
2 GB Vram GDDR5

One of threads in the forum had us debating this. I think it was the one where they showed the turbo and electric draw of the graphics card. Basically the d500 is a severely underclocked tahti (7970hd core). And the d300 is a picturn (*not sure how to spell it) (7870 core), it is also more overclocked and has a higher turbo boost than the d500.

 

[[G5]Hydra]

Macworld tested the 8-core D700 equipped MP with Unigine Heaven:
http://www.macworld.com/article/2082568/lab-tested-new-mac-pro-is-the-speedster-weve-been-waiting-for-finally.html?page=2

I couldn't find the settings but:
1280x720 103.70 fps
1920x1080 31.50 fps
2560x1600 14.40 fps

you figure 1600x900 should be somewhere around 40 fps?

 

[Anim]

One of threads in the forum had us debating this. I think it was the one where they showed the turbo and electric draw of the graphics card. Basically the d500 is a severely underclocked tahti (7970hd core). And the d300 is a picturn (*not sure how to spell it) (7870 core), it is also more overclocked and has a higher turbo boost than the d500.

Ah ok, thanks. Will have a search on the forums. It is still odd to be beaten when somebody upgraded the GPU from the stock model unless the card excels elsewhere, otherwise it needs fixing. I guess Apple could software mod the clock rate. Lots of D500 owners are going to be shouting at Apple for misleading them if it holds true in other tests.

 

[666sheep]

Lots of D500 owners are going to be shouting at Apple for misleading them if it holds true in other tests.

Only if gaming was the main reason why they have chosen D500. In all other task D500 is faster, especially in GPGPU. I doubt that there will be many such users

 

[paulrbeers]

Lots of D500 owners are going to be shouting at Apple for misleading them if it holds true in other tests.

Sure, if Apple had touted the Mac Pro as a gaming machine....

Last I checked they referred to this as a workstation.

 

[Anim]

Sure, if Apple had touted the Mac Pro as a gaming machine....

Last I checked they referred to this as a workstation.

What nonsense, What do you think professional game developers work on? Thats 3D Artists, Animators, Audio production engineers, programmers etc. They use the the best machine for the job and that would be workstations in a professional environment. A workstation is regarded as stability, reliability and support, thats it.

 

[2128506]

What nonsense, What do you think professional game developers work on? Thats 3D Artists, Animators, Audio production engineers, programmers etc.

And for most of them Mac Pro would be quite nice machine. Especially for 3D/animation.

Gameplay/performance testing is done on consumer-level cards and drivers, not on FirePros.

 

[Anim]

Only if gaming was the main reason why they have chosen D500. In all other task D500 is faster, especially in GPGPU. I doubt that there will be many such users

Aye, but there are many industries that require the same "gaming" performance.

Virtual Realty Simulations (uses game engines like Unity3D and Unreal engine)
Medical Analysis (again, fast animation showing slices or Stereoscopic 3D)
Cad Engineers (animated stress tests, OpenGL viewport performance)

See what I mean?

People assume 3D benchmarking just covers games. It doesn't

 

[paulrbeers]

What nonsense, What do you think professional game developers work on? Thats 3D Artists, Animators, Audio production engineers, programmers etc. They use the the best machine for the job and that would be workstations in a professional environment. A workstation is regarded as stability, reliability and support, thats it.

And what utter non-sense by you. Considering those who use Mari have already stated that the Mac Pro is the fastest workstation they've ever used, and programs like Maya is mostly CPU based (with only limited CUDA support) I'd say 3D Animation would be better than an iMac wouldn't you? Also Audio and Programmers aren't going to give two hoots about OpenGL. The only ones who care about FPS of OpenGL are the testers. For them an iMac or PC would make more sense would it not (especially since stability is of less concern).

Again, Apple never called this a gaming machine. They called it a work station. We've already covered the fact that for WORK it is a thing of beauty and power.

----------

Aye, but there are many industries that require the same "gaming" performance.

Virtual Realty Simulations (uses game engines like Unity3D and Unreal engine)
Medical Analysis (again, fast animation showing slices or Stereoscopic 3D)
Cad Engineers (animated stress tests, OpenGL viewport performance)

See what I mean?

People assume 3D benchmarking just covers games. It doesn't

And for all of those uses, the OpenGL use will be more than adequate. It isn't like scoring 40 in the Heaven benchmark is pathetic (oh yeah and most likely the benchmark is currently only utilizing 1 GPU since Crossfire isn't supported on OSX). And the rest of the time they aren't doing those specific tasks, they will have a killer machine that does the rest of their work as fast as possible.

 

[Anim]

And what utter non-sense by you. Considering those who use Mari have already stated that the Mac Pro is the fastest workstation they've ever used, and programs like Maya is mostly CPU based (with only limited CUDA support) I'd say 3D Animation would be better than an iMac wouldn't you?

Eh? Where did this come from. Are you mixing my response with somebody else's somehow? I know first hand all about 3D modelling an animation. I do it myself but your response has nothing to do with my response

Also Audio and Programmers aren't going to give two hoots about OpenGL. The only ones who care about FPS of OpenGL are the testers. For them an iMac or PC would make more sense would it not (especially since stability is of less concern).

Yes. testers use low end, mid and high end gaming rigs with all manner of mixed hardware and drivers to test the builds on. What I was saying is that the builds are done on workstations before it goes to testers and the developers test themselves, on the workstation they are working on.

Again, Apple never called this a gaming machine. They called it a work station. We've already covered the fact that for WORK it is a thing of beauty and power.

I already described what a workstation is, lets just agree to disagree here. I didn't call it a gaming machine either btw.

 

[mrsavage1]

I actually created a thread called "Skip the D500?". I thought people should be aware that paying a few hundred more might not get them the gaming performance they want. I am at a crossroads on spending cash for a d300 or getting a d700. I don't think the d500 is an option if you want to game. But 1k more for the d700 is really something seriously steep for a better gaming gpu

 

[Anim]

I actually created a thread called "Skip the D500?". I thought people should be aware that paying a few hundred more might not get them the gaming performance they want. I am at a crossroads on spending cash for a d300 or getting a d700. I don't think the d500 is an option if you want to game. But 1k more for the d700 is really something seriously steep for a better gaming gpu

Edited:

Totally agree with you. But if you spend that $1000 on a D700 then you have a muti-function machine that does it all, work + gaming. If you spent that $1000 on building a PC based gaming machine, would it perform faster than your D700 nMP? Probably not considering all the other high end components. It might run more games though as I am sure you would go Nvidia for the drivers.

 

[666sheep]

Virtual Realty Simulations (uses game engines like Unity3D and Unreal engine)
Medical Analysis (again, fast animation showing slices or Stereoscopic 3D)
Cad Engineers (animated stress tests, OpenGL viewport performance)

Sure, but aren't they rely on benchmarks more relevant to their workflow or 3D engine they use? I.e. SpecPerf in case of CAD people. Unigine does not seem to be particularly popular in areas other than gaming.

 

[Quash]

Looking at the 7970 vs 7870 which seems reasonable to me, it seems the D700 should get around 30% better frame rates. Very much dependent on the game.

If that's worth 1000$ dollars to you. I think i rather turn some settings down one notch. I mostly play blizzard games anyway, D300 should handle those just fine i think.

 

[iPhysicist]

@ Anim besides the non gaming GPU inside the nMP

please explain to the common non gamer (except RR3 on the iPad) why I would need ECC Ram and Workstation Class CPU to build a gaming machine?


Edit:

Medical Analysis (again, fast animation showing slices or Stereoscopic 3D)
Cad Engineers (animated stress tests, OpenGL viewport performance)

As far as i know these are not real time, at least the last one I know for sure.

 

[avemestr]

Macworld tested the 8-core D700 equipped MP with Unigine Heaven:
http://www.macworld.com/article/2082568/lab-tested-new-mac-pro-is-the-speedster-weve-been-waiting-for-finally.html?page=2

I couldn't find the settings but:
1280x720 103.70 fps
1920x1080 31.50 fps
2560x1600 14.40 fps

you figure 1600x900 should be somewhere around 40 fps?

Those numbers are meaningless, since they don't specify other settings than resolution. AA, tessellation, etc. can have a huge impact on performance.

The benchmarks and comparisons that MacWorld made are simply hilarious. Not in the funny, but in the stupid way.

 

[wildmac]

Ok, what is it in these specs that is making the D500 such a dog for regular tasks?

Is it just the lower clock speeds? If so, is the D700 going to be equally as bad for normal loads?

Or is it the lower clock combined with the lower percision ratings, where the D700 rates much better?