Dx 11 benchhmark for ATI guys

I ran it in D3D11 mode on my GTX275 as well, not all DirectX11 features need DX11 hardware. It probably just ran through like normal but without tessellation.
 
Runs pretty good on my 4870x2, Both GPU is pegged at 95+%.

Losing tessellation makes a lot of it look like shit though. Instead of stairs you get a slope. No to mention the dragon.

FPS: 54.2 Scores:1366

Render: direct3d11
Mode: 1920x1200 fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy:4x
Occlusion: enabled
Refraction: enabled
Volumetric:enabled
 
Thats very high score for 4870x2 you sure you didn't miss any setting? i got same score with my sig rig?
 
psyside said:
Thats very high score for 4870x2 you sure you didn't miss any setting? i got same score with my sig rig?
Click to expand...

Tessellation is a huge performance hit. Disable tessellation and your score should be far higher than his.

Whether or not that explains the whole difference is another matter.
 
Andrmgic said:
Tessellation is a huge performance hit. Disable tessellation and your score should be far higher than his.

Whether or not that explains the whole difference is another matter.
Click to expand...

No AA. 4AF, standard AI Cat 9.10 W7

Definitely tessellation making the difference. The difference in polygon count is pretty damn big. Like I said the stairs completely disappear and just become a slope. I'll post shots later.
 
Here's a wire frame shot, pay no attention to the fps as my GPU are greatly underclocked in 2D mode to save heat.

From the shot you can see that the 4870X2 is working with vastly less detail.
If I didn't tell that you there are supposed to be a 3 step stair leading up to the platform the dragon you wouldn't have noticed.
72451825.png
 
Last edited:
Unigine
Heaven Demo v1.0
FPS:
27.8
Scores:
701
Hardware
Binary:
Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system:
Windows 7 (build 7600) 64bit
CPU model:
Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz
CPU flags:
3399MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT
GPU model:
ATI Radeon HD 5800 Series 8.661.0.0 1024Mb
Settings
Render:
direct3d11
Mode:
1920x1200 4xAA fullscreen
Shaders:
high
Textures:
high
Filter:
trilinear
Anisotropy:
4x
Occlusion:
enabled
Refraction:
enabled
Volumetric:
enabled
Unigine Corp. © 2005-2009


cpu overclocked as described, stock GPU settings on a 5870
 
SonDa5 said:
Sweet.

What resolution did you run the test at?



Try the OpenGl mode.


On my OC HD5850 this is how it measures up.

1stheavenBenchmarkHD5850.jpg
Click to expand...

Here is my 5850 score @ your settings, we get pretty much the same score it seems. :cool:

unibmark.jpg
 
So,from what I see so far,the Nvidia cards actually perform better than the ATI cards in DX11??
 
Yes, Nvidia cards are faster since they aren't running dx11
 
g1xx3r said:
So,from what I see so far,the Nvidia cards actually perform better than the ATI cards in DX11??
Click to expand...

No, that's not what's happening. The program and/or a combination of the video card/driver attempts to run DX11 mode with all the bells and wistles aka top of the line graphics features. The attempt fails when tessolation or some feature is initialized. The simulation then defaults down to DX10 mode without any special graphics. The benchmark given although it says DX11 mode is actually a benchmark for the DX10 mode since its defaulted to the lower rendering.

Its the same way, you can sometimes find DX9 games run faster than DX10 versions, but, you lose the eye-candy. When running in DX11 mode versus DX10 mode, you lose some FPS but get a better looking picture/IQ(Image Quality). The nVidia cards can't do DX11 so they default into DX10 mode which is faster than DX11 mode but looks less beautiful/more ugly.

The best analogy I can think of is an ATI carrd and nVidia card decide to go for a off-road street race. The nVidia card doesn't have off road tires or suspension so he ops to take the city bus to where the race finishes. The city bus arrives first but the ATI card was a much more fun/interesting experience and only the people with the ATI cards actually competed in the race.
 
PS3 said:
Yes, Nvidia cards are faster since they aren't running dx11
Click to expand...

I just reran the test in DX10 and DX11 and the scores are pretty much the same so you may be right..Also even in"DX11"mode my Logi G15 shows D3D10.I thought DX10 cards were capable of running DX11 just not with all the bells and whistles?
 
g1xx3r said:
I just reran the test in DX10 and DX11 and the scores are pretty much the same so you may be right..Also even in"DX11"mode my Logi G15 shows D3D10.I thought DX10 cards were capable of running DX11 just not with all the bells and whistles?
Click to expand...

I think I may be right because you don't have a dx11 card, and this benchmark is purposely built for dx11. Also you are running gtx285 sli, so that is a killer setup...please don't compare your set up to my little 5850 :p
 
g1xx3r said:
I just reran the test in DX10 and DX11 and the scores are pretty much the same so you may be right..Also even in"DX11"mode my Logi G15 shows D3D10.I thought DX10 cards were capable of running DX11 just not with all the bells and whistles?
Click to expand...

dx10 cant do tesselation, which is causing you to gain so much fps.
 
PS3 said:
I think I may be right because you don't have a dx11 card, and this benchmark is purposely built for dx11. Also you are running gtx285 sli, so that is a killer setup...please don't compare your set up to my little 5850 :p
Click to expand...

lol..I noticed the stairs in the benchmark are sloped in both modes were I assume in DX11 they look like..well..stairs..:)
 
WBurchnall said:
No, that's not what's happening. The program and/or a combination of the video card/driver attempts to run DX11 mode with all the bells and wistles aka top of the line graphics features. The attempt fails when tessolation or some feature is initialized. The simulation then defaults down to DX10 mode without any special graphics. The benchmark given although it says DX11 mode is actually a benchmark for the DX10 mode since its defaulted to the lower rendering.

Its the same way, you can sometimes find DX9 games run faster than DX10 versions, but, you lose the eye-candy. When running in DX11 mode versus DX10 mode, you lose some FPS but get a better looking picture/IQ(Image Quality). The nVidia cards can't do DX11 so they default into DX10 mode which is faster than DX11 mode but looks less beautiful/more ugly.

The best analogy I can think of is an ATI carrd and nVidia card decide to go for a off-road street race. The nVidia card doesn't have off road tires or suspension so he ops to take the city bus to where the race finishes. The city bus arrives first but the ATI card was a much more fun/interesting experience and only the people with the ATI cards actually competed in the race.
Click to expand...

I got one.

Let's say the ATI card and Nvidia card go on a date at a pizza bar. Nvidia Card orders a couple of ready made pizza slices while the ATI card orders Steak and Eggs. Since the pizza slices are already made(DX10), of course the Nvidia card is going to be faster and eat his meal first. In the mean time, the ATI card must wait for the Steak and Eggs(DX11 effects) as it requires more time to cook.
 
g1xx3r said:
lol..I noticed the stairs in the benchmark are sloped in both modes were I assume in DX11 they look like..well..stairs..:)
Click to expand...

Lol, that's my guess. Most people probably won't be able to tell the difference between dx10 and dx11. And someone let me know what tesellation is exactly, that way I know what to look for :D
 
g1xx3r said:
I just reran the test in DX10 and DX11 and the scores are pretty much the same so you may be right..Also even in"DX11"mode my Logi G15 shows D3D10.I thought DX10 cards were capable of running DX11 just not with all the bells and whistles?
Click to expand...

You thought wrong

The Radeon HD 5000 series is the first generation of cards to support DX11
 
PS3 said:
Lol, that's my guess. Most people probably won't be able to tell the difference between dx10 and dx11. And someone let me know what tesellation is exactly, that way I know what to look for :D
Click to expand...

DX11 can just be used for performance improvements, without adding increased visuals. All up to the developer to decide that.
 
Brent_Justice said:
DX11 can just be used for performance improvements, without adding increased visuals. All up to the developer to decide that.
Click to expand...

I'd be more interested in the performance improvements. I usually have found at least from dx9 to dx10, that the difference in image quality doesn't warrant the performance hit.
 
Man I ran the benchmark on my 5870. It just kills my poor gpu.. looks like I have to divert some funds (wife patrol) for that 2nd 5870 for crossfireX
 
Ran on Vista, w/ DirectX 11 Beta installed, stock ATI 5870.

All default except for the resolution: changed to 1920x1200:

Capture.JPG


00020.jpg


Time: 253.903
Frames: 9164
FPS: 36.0925
Scores: 909.169

System: Windows Vista (build 6002, Service Pack 2) 64bit
CPU: Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz 3400MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT [Actually 3.4GHz]
GPU: ATI Radeon HD 5800 Series 8.01.01.961
System memory: 4093 Mb
Video memory: 1024 Mb
---- MathLib ----
Set SSE3 simd processor
---- Sound ----
NULL
---- Render ----
Renderer: ATI R800 1024Mb
Direct3D11 desc: ATI Radeon HD 5800 Series
Found feature level 11.0
Found compute shader 11.0
Maximum texture size: 16384
Maximum render targets: 8
---- Physics ----
Physics: Multi-threaded

Oh Yeah, pretty cool video/demo/benchmark........very nice graphics. Pretty smooth w/ the 5870, but this may vary w/ settings.


.
 
Last edited:
PS3 said:
Here is my 5850 score @ your settings, we get pretty much the same score it seems. :cool:

unibmark.jpg
Click to expand...



We have different series numbers. I think that is for the drivers. I'm running the ATI beta 9.11 drivers.


Also my Q9550 was going 4.3 ghz. This benchmark seems to be GPU dependent since it doesn't seem to take the speed of the CPU into account much.


Other big thing that I didn't mention was that I ran my benchmark with Super Sampling AA on.


I'm going to see how tweaking the CC 3d settings will help my score out. I have to beat you. :D
 
rig in sig

FPS:
22.4
Scores:
564

Render:
direct3d11
Mode:
1920x1200 8xAA fullscreen
Shaders:
high
Textures:
high
Filter:
trilinear
Anisotropy:
16x
Occlusion:
enabled
Refraction:
enabled
Volumetric:
enabled
 
Brent_Justice said:
You thought wrong

The Radeon HD 5000 series is the first generation of cards to support DX11
Click to expand...

So really the only thing Im missing out on is the tesselation?Ive seen real stairs in DX9 mode games.:p..Im sure Dirt 2 will show of some stuff but what will the performance be like..Im spoiled @ 60fps..
 
SonDa5 said:
We have different series numbers. I think that is for the drivers. I'm running the ATI beta 9.11 drivers.


Also my Q9550 was going 4.3 ghz. This benchmark seems to be GPU dependent since it doesn't seem to take the speed of the CPU into account much.


Other big thing that I didn't mention was that I ran my benchmark with Super Sampling AA on.


I'm going to see how tweaking the CC 3d settings will help my score out. I have to beat you. :D
Click to expand...

No you don't have to beat me, this is not a contest... :p

Oh and it doesn't take cpu in account much because I was running at 3.6ghz ;)
 
Nice benchy, done with 0AA, 2AA, 4AA and 8AA 1680x1050 res son's comp sig.

Unigine
Heaven Demo v1.0
FPS: 54.2
Scores: 1366

Hardware
Binary: Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system: Windows 7 (build 7100) 64bit
CPU model: AMD Phenom(tm) II X2 550 Processor
CPU flags: 3100MHz MMX+ 3DNow!+ SSE SSE2 SSE3 SSE4A HTT
GPU model: ATI Radeon HD 4770 8.661.0.0 512Mb

Settings
Render: direct3d11
Mode: 1680x1050 fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled


Unigine
Heaven Demo v1.0
FPS: 47.0
Scores: 1184

Hardware
Binary: Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system: Windows 7 (build 7100) 64bit
CPU model: AMD Phenom(tm) II X2 550 Processor
CPU flags: 3100MHz MMX+ 3DNow!+ SSE SSE2 SSE3 SSE4A HTT
GPU model: ATI Radeon HD 4770 8.661.0.0 512Mb

Settings
Render: direct3d11
Mode: 1680x1050 2xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled


Unigine
Heaven Demo v1.0
FPS: 39.2
Scores: 987

Hardware
Binary: Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system: Windows 7 (build 7100) 64bit
CPU model: AMD Phenom(tm) II X2 550 Processor
CPU flags: 3100MHz MMX+ 3DNow!+ SSE SSE2 SSE3 SSE4A HTT
GPU model: ATI Radeon HD 4770 8.661.0.0 512Mb

Settings
Render: direct3d11
Mode: 1680x1050 4xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled


Unigine
Heaven Demo v1.0
FPS: 13.3
Scores: 334

Hardware
Binary: Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system: Windows 7 (build 7100) 64bit
CPU model: AMD Phenom(tm) II X2 550 Processor
CPU flags: 3100MHz MMX+ 3DNow!+ SSE SSE2 SSE3 SSE4A HTT
GPU model: ATI Radeon HD 4770 8.661.0.0 512Mb

Settings
Render: direct3d11
Mode: 1680x1050 8xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
 
Last edited:
g1xx3r said:
So really the only thing Im missing out on is the tesselation?Ive seen real stairs in DX9 mode games.:p..Im sure Dirt 2 will show of some stuff but what will the performance be like..Im spoiled @ 60fps..
Click to expand...

go watch the video on unigines website and you will see just what the difference is with tesselation on and on in the benchmark.. if you want to see what tesselation will do in dirt 2.. check this video out..

links coming.. my FF keeps screwing up and wont let me paste the link..

arg its refusing to let me paste it because of some stupid hot key bullcrap in firefox and the forum..

oh well.. restarted FF 3 times and still didnt work.. guess you just have to search for it.. but its on techpowerup under the trailers videos
 
Last edited:
DX 10 mode -

Unigine
Heaven Demo v1.0
FPS:
43.8
Scores:
1103
Hardware
Binary:
Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system:
Windows 7 (build 7600) 64bit
CPU model:
Intel(R) Core(TM)2 Duo CPU E8200 @ 2.66GHz
CPU flags:
3200MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT
GPU model:
ATI Radeon HD 5800 Series 8.661.0.0 1024Mb
Settings
Render:
direct3d10
Mode:
1920x1080 4xAA fullscreen
Shaders:
high
Textures:
high
Filter:
trilinear
Anisotropy:
16x
Occlusion:
enabled
Refraction:
enabled
Volumetric:
enabled


DX 11 Mode -

Unigine
Heaven Demo v1.0
FPS:
28.6
Scores:
720
Hardware
Binary:
Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system:
Windows 7 (build 7600) 64bit
CPU model:
Intel(R) Core(TM)2 Duo CPU E8200 @ 2.66GHz
CPU flags:
3200MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT
GPU model:
ATI Radeon HD 5800 Series 8.661.0.0 1024Mb
Settings
Render:
direct3d11
Mode:
1920x1080 4xAA fullscreen
Shaders:
high
Textures:
high
Filter:
trilinear
Anisotropy:
16x
Occlusion:
enabled
Refraction:
enabled
Volumetric:
enabled

Dx 11 Mode w/o AA -

Unigine
Heaven Demo v1.0
FPS:
37.4
Scores:
941
Hardware
Binary:
Windows 32bit Visual C++ 1500 Release Oct 22 2009
Operating system:
Windows 7 (build 7600) 64bit
CPU model:
Intel(R) Core(TM)2 Duo CPU E8200 @ 2.66GHz
CPU flags:
3200MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 HTT
GPU model:
ATI Radeon HD 5800 Series 8.661.0.0 1024Mb
Settings
Render:
direct3d11
Mode:
1920x1080 fullscreen
Shaders:
high
Textures:
high
Filter:
trilinear
Anisotropy:
16x
Occlusion:
enabled
Refraction:
enabled
Volumetric:
enabled
 
PS3 said:
I got one.

Let's say the ATI card and Nvidia card go on a date at a pizza bar. Nvidia Card orders a couple of ready made pizza slices while the ATI card orders Steak and Eggs. Since the pizza slices are already made(DX10), of course the Nvidia card is going to be faster and eat his meal first. In the mean time, the ATI card must wait for the Steak and Eggs(DX11 effects) as it requires more time to cook.
Click to expand...

oh the old steak and eggs analogy? thats all you had to say...
 
sirmonkey1985 said:
go watch the video on unigines website and you will see just what the difference is with tesselation on and on in the benchmark.. if you want to see what tesselation will do in dirt 2.. check this video out..

links coming.. my FF keeps screwing up and wont let me paste the link..

arg its refusing to let me paste it because of some stupid hot key bullcrap in firefox and the forum..

oh well.. restarted FF 3 times and still didnt work.. guess you just have to search for it.. but its on techpowerup under the trailers videos
Click to expand...

link http://www.youtube.com/watch?v=bkKtY2G3FbU
 
I have a Sapphire 5870 and the benchmark does not work...it installed fine but every time I try to run the benchmark I keep getting this error...I tried downloading the latest version of OpenAL and it still would not work...anyone have any ideas?

errorcz.jpg
 
Last edited:
You must log in or register to reply here.
Back
Top