Unigine Superposition Benchmark - post your results!

[jboydgolfer]

i mustve done something wrong, my results arent in list in OP

 

[Fluffmeister]

Fluff, I think you have your memory and core clocks switched in your post.

You had me doubling checking for a sec! But yeah it's right, the older GDDR5 runs at a higher clock, these are stock clocks on my Palit Super JetStream but she does boost to 1354 of her own accord.

 

[Slizzo]

You had me doubling checking for a sec! But yeah it's right, the older GDDR5 runs at a higher clock, these are stock clocks on my Palit Super JetStream but she does boost to 1354 of her own accord.

View attachment 86872

Huh, I thought most people had their cards boost quite a bit higher than what you have listed. Not bad if that just a normal boost clock.

 

[Kanan]

Look at the settings for the extreme preset as compared to the settings for the 4K preset. The extreme preset (regardless of resolution) has the maximum options enabled , and is therefore significantly more strenuous on both gpu & vRAM. Even with a brand-new RX 580 scores are less than 3000 on the extreme preset. It makes me wonder what was AMD thinking? Their "top of the line" GPU is 100%+ less powerful than Nvidia's reference 1080. Of course it's also half the cost or even more like a third.personally I always loved AMD GPUs, I just wish they could get back in the game with a single chip video card that was kick ass

"Thinking" doesn't invent you faster GPUs, but Vega is coming.

[updated score]

Single GPU - 1080p Extreme

3482 | uuuaaaaaa | Phenom II X6 1100T @ 3.6GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 500 MHz

View attachment 86848

Asus R9 Fury Strix, 4096 shaders, 375W/325A bios. Cooled by a Thermalright Shaman (custom bracket) + VRM heatsink milled out of a ram heatpipe cooler.

Try HBM overclocking if you want a higher score. I'd also like to see it for the funs.

7851 - xkm1948 / 5820K 4.25 GHz / R9 FuryX @ 1100 MHz (actual boost clock) / HBM 500 MHz

View attachment 86851

4802 - xkm1948 / 5820K 4.25 GHz / R9 FuryX @ 1100 MHz (actual boost clock) / HBM 500 MHz

View attachment 86853

Same here, try HBM overclocking, I would like to see the difference.

13726|FireWallCR|I7 3960x 4.2 GHz|1GamingX 1070 @ ACTUAL STOCK MHz (actual boost clock)

I need GPU and VRAM clocks. Use Afterburner or smth similar to see them.

Still.. Almost an 8 year old chip keeping up with most newer stuff! I wish I would of kept mine.

Yeah, and those are kinda living on in consoles atm, still doing relatively strong @ max. 60 FPS gaming.

It easily does 4.0GHz (20x multi) and 2800MHz NB, too bad that I'm stuck with this crappy MSI board, my Crosshair III died last year :/

Ah, shit! That was a kickass Phenom II setup you had. Still is nice enough. How is playing Battlefield 4 or 1 with it?

Full gamut of runs again, this time with a GTX 1080 Ti.

1080P Medium
20848 | Slizzo | i7 6800K 4.2GHz | GTX 1080 Ti @ 2050 MHz |1502 MHz


1080P High
16630 | Slizzo | i7 6800K 4.2GHz | GTX 1080 Ti @ 2050 MHz |1502 MHz


1080P Extreme
6171 | Slizzo | i7 6800K 4.2GHz | GTX 1080 TI @ 2050 MHz |1502 MHz


4K Optimized
10167 | Slizzo | i7 6800K 4.2GHz | GTX 1080 Ti @ 2050 MHz |1502 MHz


8K Optimized
4555 | Slizzo | i7 6800K 4.2GHz | GTX 1080 TI @ 2037 MHz |1502 MHz

You're first on 1080P High now. Seems that for 1080P Medium your clocks are too low and limit the GPU a bit but for 1080P High it seems to favour high L3 cache and doesn't care as much about the clocks too. /speculation

It's painful following Slizzo, but hey ho, my 8 year old 920 with a 980 Ti shaped cherry on top.

1080P Medium
12509 | Fluffmeister | i7 920 @ 2.6 GHz | Palit GTX 980 Ti @ 1354 MHz | 1753 MHz

1080P High
9195 | Fluffmeister | i7 920 @ 2.6 GHz | Palit GTX 980 Ti @ 1354 MHz | 1753 MHz

1080P Extreme
3731 | Fluffmeister | i7 920 @ 2.6 GHz | Palit GTX 980 Ti @ 1354 MHz | 1753 MHz

4K Optimized
5682 | Fluffmeister | i7 920 @ 2.6 GHz | Palit GTX 980 Ti @ 1354 MHz | 1753 MHz

You should overclock that nice CPU, it's a pain to see it running on stock!

i mustve done something wrong, my results arent in list in OP

Which ones do you mean? Maybe I responded to you and told you smth is wrong, or I didn't see it.

---

Updated!

 

[uuuaaaaaa]

Try HBM overclocking if you want a higher score. I'd also like to see it for the funs.

Single GPU - 1080p Extreme

HBM 400 MHz
3470 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 400 MHz

HBM 500 MHz
3482 | uuuaaaaaa | Phenom II X6 1100T @ 3.6GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 500 MHz

HBM 545 MHz
3472 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 545 MHz

HBM 600 MHz
3468 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 600 MHz

Negative scaling with HBM overlock, this is nothing new. Some applications scale positively with HBM clock like geekbench compute:

MY score @ 1226MHz/ HBM 600 MHz
http://hwbot.org/submission/3383005_

 

[Kanan]

Single GPU - 1080p Extreme

HBM 400 MHz
3470 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 400 MHz

HBM 500 MHz
3482 | uuuaaaaaa | Phenom II X6 1100T @ 3.6GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 500 MHz

HBM 545 MHz
3472 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 545 MHz

HBM 600 MHz
3468 | uuuaaaaaa | Phenom II X6 1100T @ 3.7GHz | Asus R9 4096 Fury Strix @ 1097 MHz | HBM 600 MHz

Negative scaling with HBM overlock, this is nothing new. Some applications scale positively with HBM clock like geekbench compute:

MY score @ 1226MHz/ HBM 600 MHz
http://hwbot.org/submission/3383005_

Thanks, I guess on higher clocks the latency increases too much, thus worse scores, because the benchmark isn't bandwidth limited - that would increase the points then, counteract the increased latency.

Did you see my question about BF4 / BF1 gaming on Phenom II 1100T? I replaced my Phenom II 940 because of BF4, but it was with DDR2 800 and the bandwidth wasn't sufficient for that heavy game anymore (still good enough for Crysis 3 for example). I guess yours does better with DDR3.

 

[uuuaaaaaa]

Thanks, I guess on higher clocks the latency increases too much, thus worse scores, because the benchmark isn't bandwidth limited - that would increase the points then, counteract the increased latency.

Did you see my question about BF4 / BF1 gaming on Phenom II 1100T? I replaced my Phenom II 940 because of BF4, but it was with DDR2 800 and the bandwidth wasn't sufficient for that heavy game anymore (still good enough for Crysis 3 for example). I guess yours does better with DDR3.

The 400MHz strap has tighter timings, that is why I tried the lower clock (HBM clocks in steps, i.e. if you set it at 510 it will run at 500, but 535 will run 545, 580 will default to 600 and so on.). I'm on DDR3 1333Mhz @6-7-6-16-33-1T, it works better than 1600MHz with looser timings. I can't test BF4/BF1 atm, I do not have any of those games :|

 

[Carrotsalesman]

13935|Carrotsalesman|CPU 3930k 4.2 GHz|GPU @ 1641 MHz| Memory clock 2336 MHz

 

[Carrotsalesman]

3848|Carrotsalesman|CPU 3930k 4.2 GHz|GPU @ 1641 MHz| Memory clock 2336 MHz

 

[Kanan]

The 400MHz strap has tighter timings, that is why I tried the lower clock (HBM clocks in steps, i.e. if you set it at 510 it will run at 500, but 535 will run 545, 580 will default to 600 and so on.).

Yeah I remember. Curious if HBM2 will be as special too, I guess yea.

Updated.

 

[infrared]

Nice benchmark, although I had to sort out the SLI Profiles for it to run on both cards..

6700K @ 4.8GHz, 4.5GHz cache
Ram @ 3333MHz 15-17-17-34 1T
2x GTX Titan (Kepler) @ 1175MHz/3400MHz

1080P Medium:
18071|infrared|6700K 4.8 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz



1080P Extreme:
5351|infrared|6700K 4.8 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

 

[CAPSLOCKSTUCK]

I am loving this thread.

Nice one @Kanan

 

[Kanan]

Nice benchmark, although I had to sort out the SLI Profiles for it to run on both cards..

6700K @ 4.8GHz, 4.5GHz cache
Ram @ 3333MHz 15-17-17-34 1T
2x GTX Titan (Kepler) @ 1175MHz/3400MHz

1080P Medium: 18071

View attachment 86894

1080P Extreme: 5351

View attachment 86895

5. Follow this format for posting (and don't forget the screenshot provided by Superposition):
SCORE|Username|CPU X.x GHz|GPU @ ACTUAL XXXX MHz (actual boost clock) | Memory clock (no double data+ rate clocks, actual real clocks) XXXX MHz

Please.
--
Thanks Ger, I'm happy it's running so well, people seem to like the bench. =)

 

[ISI300]

Gave it a go on my trusty old 780 Ti as well as my trustier and older 580 (just for funsies).
GTX 780 Ti @ Stock
2320|ISI300|i5 4670K 3.4 GHz|780 Ti @ 993| Memory clock 1750MHz

GTX 580 @ Stock (772/1544/1002)
3627|ISI300|i5 4670K 3.4 GHz|GTX 580 @ 772 MHz|Memory clock 1002 MHz|

Edited to follow rule #5 correctly

 

[infrared]

Please.

Sorry! Post above edited. Thanks for keeping score

 

[Carrotsalesman]

Hey Kanan,
Keep up the good work! Didn't think i'd be 10th overall on 1080p medium, nowhere near! Top idea you had mate.
Final post is my "high" setting score, Thanks.

10062|Carrotsalesman|CPU 3930k 4.2 GHz|GPU @ 1641 MHz| Memory clock 2336 MHz

 

[Kanan]

Sorry! Post above edited. Thanks for keeping score

Np. =) I inserted your scores for 1080p Extreme, no 1080p Medium @ MGPU because of low demand/low use.

--
Updated.

 

[infrared]

Np. =) I inserted your scores for 1080p Extreme, no 1080p Medium @ MGPU because of low demand/low use.

--
Updated.

That's cool, cheers. I did a 1080P high run with MGPU too

13287|infrared|i7 6700K 4.8 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

 

[Carrotsalesman]

2685|Carrotsalesman|CPU 3930k 4.2 GHz|GPU @ 1641 MHz| Memory clock 2336 MHz

 

[Kanan]

That's cool, cheers. I did a 1080P high run with MGPU too

13287|infrared|i7 6700K 4.8 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

View attachment 86902

Cool, that's a really high score. If you like do 4K & 8K as well, let's put that 6 GB vram to use!
---
Updated.

 

[infrared]

Cool, that's a really high score. If you like do 4K & 8K as well, let's put that 6 GB vram to use!
---
Updated.

Niiice, I didn't realize it could run at higher than native resolution, that's a really cool feature!

4K ran nicely, but 8K bought it to it's knees lol

7923|infrared|6700K 4.7 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz



3172|infrared|6700K 4.7 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

 

[basco]

maybe dumm question but what means dpi scaling in the options?(i know what it means in windows)
is it for the 4k\8k benches to scale down on full hd monitors?
TIA

 

[jboydgolfer]

Which ones do you mean? Maybe I responded to you and told you smth is wrong, or I didn't see it.

https://www.techpowerup.com/forums/...chmark-post-your-results.232298/#post-3637222

It looks right, i posted on 1st page

 

[Kanan]

Niiice, I didn't realize it could run at higher than native resolution, that's a really cool feature!

4K ran nicely, but 8K bought it to it's knees lol

7923|infrared|6700K 4.7 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

View attachment 86907

3172|infrared|6700K 4.7 GHz|2x GTX Titan (6GB Kepler) @ 1175 MHz | 1700 MHz

View attachment 86908

Still very nice scores =)

https://www.techpowerup.com/forums/...chmark-post-your-results.232298/#post-3637222

It looks right, i posted on 1st page

Yeah the formating is a mess, please fix.
--
Updated.

 

[Swamp Monster]

So here are my 4K results. Why not, since it runs better than Extreme anyway.
2303|Swamp Monster|Phenom II X4 940 AM2 @ 3.5 GHz|MSI R9 380 4GB @ 980 MHz | 1425 MHz
4K OPTIMIZED:

Older system deserves some love too
1340|Swamp Monster|Phenom II X4 925 AM3 @ 2.8 GHz|XFX HD 5770 1GB @ 850 MHz | 1200 MHz
1080P MEDIUM:

and another Bulldozer build
5868|Swamp Monster|FX-8320 @ 3.5 GHz|MSI R9 280 3GB @ 1150 MHz | 1550 MHz
1080P MEDIUM:

4214|Swamp Monster|FX-8320 @ 3.5 GHz|MSI R9 280 3GB @ 1150 MHz | 1550 MHz
1080P HIGH:

1769|Swamp Monster|FX-8320 @ 3.5 GHz|MSI R9 280 3GB @ 1150 MHz | 1550 MHz
1080P EXTREME: