New Performance Benchmark from Unigine

Revdarian

2[H]4U
Joined
Aug 16, 2010
Messages
2,616
Last edited:
Don't have a reference good or bad figure just post it here.

Superposition.png
 
This shit is legit. It tortures my card and ramps it up like furmark. Serisouly stressful and very realistic bench. My card never goes in the 70s but this bench takes it up to 75ish.
 
Tested this on my store's watercooled 1080+7700k @1440p on the extreme settings: brought it to its knees: 15FPS average. Ouch. This sh** is fo reeeeal.

Going to try it on my sig rig once I get home.
 
2302 at 1440p with Extreme/High settings. 980 Ti and a 6600K @ 4.6 GHz. Average fps 17.

Interestingly it runs about 50-60 fps in the Game mode with High shaders but even more interestingly, turning on 4K optimized shaders gives me about 100 fps in 1440p! At 4K that same setting gives me about 50 fps. You'd think the 4K optimized would be heavier than the High setting but I guess not.

Can't really tell what is the real difference in the Extreme setting.
 
Last edited:
2302 at 1440p with Extreme/High settings. 980 Ti and a 6600K @ 4.6 GHz. Average fps 17.

Interestingly it runs about 50-60 fps in the Game mode with High shaders but even more interestingly, turning on 4K optimized shaders gives me about 100 fps in 1440p! At 4K that same setting gives me about 50 fps. You'd think the 4K optimized would be heavier than the High setting but I guess not.

Can't really tell what is the real difference in the Extreme setting.

I clearly saw the image quality difference between extreme and high. Extreme got me 17 fps on my rx 470 machine at 1080p while high got me in the 40s. Image quality difference was clearly there.
 
Superposition dx ext.png

Superposition ogl ext.png

My scores seem a tiny bit low to me for some reason.
I'm paranoid and ocd so I'll probably just reformat soon. lol
 
Hope more people also post their open gl results just for comparison. Open gl is the ultimate stress since dx seems to run about 10-15% faster.
 
Mine with 8350 @ 4.74Ghz 290@ 1100/1300
Superposition_Benchmark_v1.0_2635_1492032595.png


Most stable bench I have ever seen, as in no wild swings.
 
6300+ ish something at the 1080p extreme preset. didnt bother to run again and i didnt snap it :D
 
The more scores I see get posted, the more convinced I am that the scores I'm getting are waaay lower than they should be. :(

~edit~
hmm maybe not, gotta do more testing

I only have a 1080 monitor, but almost forgot I can run at 4k scaled due to DSR so here are my 4K results.

Superposition dx 4k.png

Superposition ogl 4k.png
 
Last edited:
I knew this was coming out, will def check it out tonight and give it a twirl. Curious as to how it will scale with my FuryX and Radeon Pro Duo.
 
Very cool benchmark (and the ability to run up to 10k resolution...)

Did some benchmarks myself. Cropped them all into a single picture to save post space. Specs are in my signature.

1111 copy.jpg
 
Last edited:
Nice new benchmark!

Only change in benchmarks I set the fan to manual 100%. 20c drop in the 4k run.

SP 1080p Medium.JPG
SP 4K.JPG
SP 8K.JPG
 
Last edited by a moderator:
1080medium.png 1080mediumogl.png 1080 high.png 1080highogl.png 1080extreme.png 1080extremeogl.png 4koptimized.png 4koptimizedogl.png 8koptimized.png

8k would crash during the loading screen in OpenGL, but was able to run it in DirectX.
 
Figured I may as well throw three screenshots in. These are with the freebie benchmark.

upload_2017-4-13_21-55-38.png
upload_2017-4-13_21-56-13.png
upload_2017-4-13_21-55-1.png
 
Only 85% GPU use.
Either an SLI problem or you are CPU limited.
Slow ram can limit cpu performance but yours is 3200MHz according to your sig, good speed.
Unfortunate.

Maybe both but benchmark at 4K optimize basically has both cards at 50%. That to me means SLI or driver limitation. At lower resolutions the usage goes up which would seem to indicate it was not a cpu restriction with the higher resolutions. In VR in general usage is around 85%-90% overall. Just some strange behavior which I do believe is driver issues with RyZen. OpenGL has both 1070's over 95% most of the time except the scores are worst. Since benchmark is DX 11 as far as I can tell, does not use Vulkan, it is also API limited or designed as well. Still very nicely done!
 
Maybe both but benchmark at 4K optimize basically has both cards at 50%. That to me means SLI or driver limitation. At lower resolutions the usage goes up which would seem to indicate it was not a cpu restriction with the higher resolutions. In VR in general usage is around 85%-90% overall. Just some strange behavior which I do believe is driver issues with RyZen. OpenGL has both 1070's over 95% most of the time except the scores are worst. Since benchmark is DX 11 as far as I can tell, does not use Vulkan, it is also API limited or designed as well. Still very nicely done!
Yeah, sounds feasible.
 
benchmarks for me. Really shows the difference with an older xeon versus newer procs.

Superposition_Benchmark_v1.0_8598_1492141172.png

Superposition_Benchmark_v1.0_13589_1492140851.png
 
Have a laugh on me.
I thought Quantum Break was rough on my rig smh.
I didn't even bother with Extreme anything, High was harsh enough , especially OpenGL.
[email protected]
670GTX 2Gb @1200/3000

Superposition_Benchmark_v1.0_5488_1492145200.png

Superposition_Benchmark_v1.0_3280_1492145613.png

Superposition_Benchmark_v1.0_4942_1492145941.png

Superposition_Benchmark_v1.0_2013_1492146256.png
 
Laugh? Nah, your rig did alright imho and showcased how well this fantastic piece of pretty code scales!
 
Decided to try the 720 low preset just for the heck of it and as I suspected, hit a cpu bottleneck. As you can see, the gpu maxed at 76%.
From my previous posts, you can see that I have run every other preset and never went below 98% until this.

Superposition_Benchmark_v1.0_26198_1492211302.png
 
Decided to try the 720 low preset just for the heck of it and as I suspected, hit a cpu bottleneck. As you can see, the gpu maxed at 76%.
From my previous posts, you can see that I have run every other preset and never went below 98% until this.
Thats great but it means nothing for general PC use.
 
Interestingly enough, the cpu didn't seem to bottleneck the gpu (as much/at all?) when I ran the low preset with opengl instead of dx. Might have to retest a few times to see if it was some kind of fluke and I'll monitor everything with hwinfo during to see if I notice any anomalies.

Superposition_Benchmark_v1.0_23805_1492211896.png
 
Last edited:
Back
Top