The new Heaven GPU benchmark

[Fighterpilot]

I just downloaded it and benchmarked my rig using the DX10 codepath and at 1680x1050.
http://www.pcgameshardware.com/aid,6980 ... /Download/
It has pretty cool graphics and you can play around with the onscreen settings while its running.
Here's how my(ageing)rig scored....lets see what ya got!

[JoJoBoy]

Could someone who has a DX11 card run this benchmark in both dx10 and dx11 to give a performance comparison?

[crazybus]

Just for the lulz

[Shining Arcanine]

It seems that I have the highest score posted so far:



I find this impressive considering that my GeForce 250 GTS is based off the G92 core, which is a die shrink of the G80.

[Krogoth]

[Fighterpilot]

Thanks for posting scores guys.
I'm hoping someone will come in with Core i7 and a HD5800 series card to see how high it scores.
For a freebie its a pretty cool little graphics benchmark.

[Flying Fox]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

[Fragnificent]

As soon as my copy of Windows 7 arrives (tonight?) I'll run this on my dual 3.6 / 5850 and post scores.

[Fighterpilot]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

Yup...check his screen resolution.

[Shining Arcanine]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

It is possible that the scores are not directly comparable across different screen resolutions. I haven't checked this though.

[Mentawl]

[Flying Fox]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

Yup...check his screen resolution.

Oops, duh. Guess I needed more coffee.

[Jigar]

I want to see Leor's score in this thread...

[danny e.]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

Yup...check his screen resolution.

Oops, duh. Guess I needed more coffee.

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

[Flying Fox]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I thought it does. FP is running at 1680x1050 while Krogoth ran 1280x1024, so Krogoth's score is much higher with a lower res and CF. Wrong?

That's why real testing like TR's articles keep things at the same res for the same graph at least. We amateurs are just running these things "for the lulz".

[danny e.]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I thought it does. FP is running at 1680x1050 while Krogoth ran 1280x1024, so Krogoth's score is much higher with a lower res and CF. Wrong?

the score should be lower for running a lower resolution. The higher score for a lower resolution makes it hard to compare scores when different people run different resolutions.

What I am basically saying is I should be able to run the benchmark at a bunch of differnet resolutions and always get the same final score because my card didnt change.

The fps would change but the score should be a reflection of the card.. with the res taken into account.
I admit that it would probably be rather hard to calculate some accurate penalty for running a lower res.

[Flying Fox]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I thought it does. FP is running at 1680x1050 while Krogoth ran 1280x1024, so Krogoth's score is much higher with a lower res and CF. Wrong?

the score should be lower for running a lower resolution. The higher score for a lower resolution makes it hard to compare scores when different people run different resolutions.

I admit that it would probably be rather hard to calculate some accurate penalty for running a lower res.

I guess I am bit dense tonight, but the score seems to be based on the fps too. So since lower res should in higher fps it makes sense the score will be higher?

[crazybus]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I'm not aware of any precedent where benchmark "score" is dependent on resolution. 3DMark is the same way, bench with a higher/lower resolution setting and you'll get a different score. Now the logic behind providing a score where the FPS value will suffice is somehow missing me.

[danny e.]

I guess I am bit dense tonight, but the score seems to be based on the fps too. So since lower res should in higher fps it makes sense the score will be higher?

I edited my post a bit above to explain what I meant a little better but I almost completely change my mind now that I think about it because I think it'd be nearly impossible if not completely impossible.

[Flying Fox]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I'm not aware of any precedent where benchmark "score" is dependent on resolution. 3DMark is the same way, bench with a higher/lower resolution setting and you'll get a different score. Now the logic behind providing a score where the FPS value will suffice is somehow missing me.

Well, there are multiple scenes used and the fps values will be different. So kind of a weighted average of the scenes to give a score?

[Blazex]

only amd cpu so far and seems it was running at 2.6 apparently...

[Dizzytaz00]

OK!! Tried the software for giggles
&
I can't get pasted 4 fps & i'm running these components:

Amd phenom II x4 955 am 3
Asus M4A78 Plus 770/700 chipset Am 2
2 wd carvier blue 500 gib sata 2 16 mb
xfx gforce 8400gs 512 mb
Ozc Reaper HCP DDR- 2 800hmz 4 gib dual channel
windows 7 64bit


Puzzled??? Can any one help?

EDIT added ram info

[Krogoth]

Krogoth's 2x4850 with a 3GHz CPU scores more than twice as Fighterpilot's 3.3GHz with 1x4850? Got an explanation?

Yup...check his screen resolution.

Oops, duh. Guess I needed more coffee.

Apparently, SLI/CF scale incredibility well in this bench. It also helped that my cards were driving 400 Kliopixels less than Fighterpilot . (1280x1024 = 1.31 Megapixels versus 1680x1050 = 1.7 Megapixels).

[Flying Fox]

OK!! Tried the software for giggles
&
I can't get pasted 4 fps & i'm running these components:

Amd phenom II x4 955 am 3
Asus M4A78 Plus 770/700 chipset Am 2
2 wd carvier blue 500 gib sata 2 16 mb
xfx gforce 8400gs 512 mb
Ozc Reaper HCP DDR- 2 800hmz 4 gib dual channel
windows 7 64bit


Puzzled??? Can any one help?

EDIT added ram info

Are you using Nvidia's latest drivers?

[mattsteg]

The fact is that just proves that the benchmark is NOT a good one because it doesn't take resolution into account for the score.

You should never get a higher score running at a lower resolution than you do running at a high resolution.

On my system I also get a much higher score running DirectX 10 than 11 with tesselation turned on

I thought it does. FP is running at 1680x1050 while Krogoth ran 1280x1024, so Krogoth's score is much higher with a lower res and CF. Wrong?

the score should be lower for running a lower resolution. The higher score for a lower resolution makes it hard to compare scores when different people run different resolutions.

What I am basically saying is I should be able to run the benchmark at a bunch of differnet resolutions and always get the same final score because my card didnt change.

The fps would change but the score should be a reflection of the card.. with the res taken into account.
I admit that it would probably be rather hard to calculate some accurate penalty for running a lower res.

You do realize that what you're asking for isn't possible for a meaningful benchmark, right?

[Airmantharp]

I don't see how one could compensate for resolution in a benchmark, at least completely. To me, if you run at a lower resolution, you're running at a higher FPS. This is taxing the GPU, yes, but the whole system is involved with the rendering of each frame. Thus, at a lower resolution, the CPU would be taxed more than at a higher resolution, which means that your end scores would still be off, if you simply account for the differences in the number of pixels to be rendered per frame.

Also, when comparing really high resolutions with much lower resolutions, such as a 30" 2560x1600 with say 1024x768, you might also get into a situation where GRAM capacity is limiting performance for the higher resolution versus the lower resolution.

Now I'm sure that the developer could 'account' for these differences by applying statistically generated adjustment curves, taking into account the range of CPU perfomance, but it still wouldn't be as accurate as benchmarking at the same resolution across the board, with the same CPU at the same settings.

[JustAnEngineer]

Unigine Heaven Demo v1.0
FPS: 15.7
Scores: 397

Hardware
Binary: Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system: Windows Vista (build 6002, Service Pack 2) 64bit
CPU model: Intel(R) Core(TM)2 Duo CPU P8700 @ 2.53GHz
CPU flags: 78MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT
GPU model: NVIDIA GeForce 9800M GS 8.16.11.8681 512Mb

Settings
Render: direct3d10
Mode: 1680x1050 fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 4x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Unigine Corp. © 2005-2009

Edit: Not too shabby for a 0.078 GHz processor?

[ChronoReverse]

Looks pretty ugly in DX10 mode in comparison to the DX11 videos I've seen. Suddenly I want a 58xx to replace my GTX260

[Airmantharp]

Looks pretty ugly in DX10 mode in comparison to the DX11 videos I've seen. Suddenly I want a 58xx to replace my GTX260

Don't fall for it! At least, not for this benchmark . I want a 5850 too, but I honestly don't have a need for it, with a GTX260 in my current system. I'll wait and see what Dragon Age, Mass Effect 2, and Starcraft 2 do to my system first.

[Dizzytaz00]

yes i'm using the lastest nivdia drivers 191