Unigine Valley Benchmark Scores Thread + SUPERPOSITION ***Over 1000 Submissions!***

[Jumper118]

Name: Da'Argo

4770K @ 4.2

PNY GTX 770 @ 1200 Boost/1978 Mem

AVG FPS: 50.8

Score: 2126

 

 

nice memory clock. will it not manage 2k?

[Lord Pantaloons]

 

themathouston

i7 3770k @ 4.7

EVGA ACX SC GTX 780 @ 1228Mhz / 1555Mhz

120.2 FPS

5030 Score

 

 

You should mention that you're running SLI, since your score is in line with other 780 SLI rigs.

[Da'Argo]

nice memory clock. will it not manage 2k?

 

nope, tried to go +20 more and got glitches.

[AznKnight82]

AznKnight82

i7-4770k @ 3.5GHz (haven't bothered to OC yet)

eVGA GTX Titan Black (+ ACX cooler) @ +165 / +300 (have no idea how to read speeds, used PrecisionX)

76.7 FPS

3207 score

 

---- I have to say, the reference cooler might be better to use than the ACX cooler in my case. The hot air from the ACX cooler has almost no where to go in my 250D (and doesn't come out of the back as much as I would like, might need a stronger fan up-front, maybe Scythe GT AP-45). Could've pushed the GPU harder, but saw that these speeds were more-or-less standard for a 780ti, so stuck with those numbers. Plus, adding just a few more MHz added a few more degrees.

 

[Mattriz]

Mattriz

i7-930 @ 2.8Ghz

Gtx 465 / 607/1603

15.9

666

Saw a used 290x for half the price of what it costs in norway, so i just had to buy it. (new 676 $, bought it for 331 $)

 

 

Mattriz

i7-930 @ 4.1Ghz

R9 290x stock / 1120/1420

65

2721

[GzeroD]

 

Saw a used 290x for half the price of what it costs in norway, so i just had to buy it. (new 676 $, bought it for 331 $)

 

Mattriz

i7-930 @ 4.1Ghz

R9 290x stock / 1120/1420

65

2721

 

Nice Congrats on that little steal.

[MegaTechPC]

I was quite proud of my little backup folding rig when I decided to see how my old GTX 580 Lightning would handle this bench.  As it turns out, pretty good!

 

Majin SSJ Eric -- Intel Core i7 2600K @ 4.8GHz -- 2 x 4GB G.Skill Ripjaws 2133MHz memory -- MSI GTX 580 Lightning 1.5GB @ 934MHz / 2264MHz -- Avg FPS 43.1 -- Score 1804

 

 

Here's the rest of the specs for my folding rig:

Corsair 600T (White)

Asus P67 Sabertooth

Corsair HX750

Corsair H80

[cnapier39428]

Ok I lied 1 more before she goes under water.

GzeroD

Intel I5 3570k @ 4.7

ASUS reference Radeon R9 290 @ 1230/1450

FPS: 69.4

SCORE: 2906

How did you get it clocked that high? You must have a golden sample because my sapphire 290 max stable clock is 1108..

[Faa]

This benchmark is here and there cpu bound especially with 2 780's. With one card is pretty much gpu bound everywhere. That run in 12/18 as far as I can remember was really cpu limited.. Turning some cores off and bumping the clock could improve your score dramatically here. Not complety understanding why the OP chose for valley instead of unigine which is a lot more gpu bound. 

[Lord Pantaloons]

This benchmark is here and there cpu bound especially with 2 780's. With one card is pretty much gpu bound everywhere. That run in 12/18 as far as I can remember was really cpu limited.. Turning some cores off and bumping the clock could improve your score dramatically here. Not complety understanding why the OP chose for valley instead of unigine which is a lot more gpu bound. 

I only saw no improvement with bumping up the clock, Valley is still very GPU bound. I ran my original score with SLI 660 Ti's with my i5 at 4.5 then I pumped it up to 4.8 for a hero run and it made no difference.

[cnapier39428]

I only saw no improvement with bumping up the clock, Valley is still very GPU bound. I ran my original score with SLI 660 Ti's with my i5 at 4.5 then I pumped it up to 4.8 for a hero run and it made no difference.

Yea I overclocked my CPU 700mhz and saw pretty much no difference myself

[Faa]

I only saw no improvement with bumping up the clock, Valley is still very GPU bound. I ran my original score with SLI 660 Ti's with my i5 at 4.5 then I pumped it up to 4.8 for a hero run and it made no difference.

I'll show you later where its cpu bound. Seeing plenty of loads like 60% only. And a 660ti in sli is nothing to feed compared to 2 780's.

[Lord Pantaloons]

I'll show you later where its cpu bound. Seeing plenty of loads like 60% only. And a 660ti in sli is nothing to feed compared to 2 780's.

It doesn't matter it shows that it's more GPU bound, going from 4.5Ghz to 4.8Ghz on a 3570k and showing showing no improvement means that it was a lack on power on the part of the GPU's.

[Faa]

Yea I overclocked my CPU 700mhz and saw pretty much no difference myself

 

Because you're only using a single R9 290. See screenshots below, 2 780's at 50% would match a 290@99%

 

It doesn't matter it shows that it's more GPU bound, going from 4.5Ghz to 4.8Ghz on a 3570k and showing showing no improvement means that it was a lack on power on the part of the GPU's.

You couldnt be more wrong: http://imgur.com/z0Tgpf2,eid7BZQ,VJn0nbZ,02U7Bsc,DXoFfh1,77SNBgp,ne9MORh,X17D24G#0

Loads around 50/50% is cpu bound. If I had a cpu that could push both to 99% there would have been a 100% fps gain.

[cnapier39428]

Because you're only using a single R9 290. See screenshots below, 2 780's at 50% would match a 290@99%

You couldnt be more wrong: http://imgur.com/z0Tgpf2,eid7BZQ,VJn0nbZ,02U7Bsc,DXoFfh1,77SNBgp,ne9MORh,X17D24G#0

Loads around 50/50% is cpu bound. If I had a cpu that could push both to 99% there would have been a 100% fps gain.

Well I actually tested it on my 290 and my 780ti but yea I see the point you're making

[ERSCreeper]

Because you're only using a single R9 290. See screenshots below, 2 780's at 50% would match a 290@99%

 

You couldnt be more wrong: http://imgur.com/z0Tgpf2,eid7BZQ,VJn0nbZ,02U7Bsc,DXoFfh1,77SNBgp,ne9MORh,X17D24G#0

Loads around 50/50% is cpu bound. If I had a cpu that could push both to 99% there would have been a 100% fps gain.

Lol what? I overclocked my 3570K from 3.4Ghz to 4.8Ghz and I haven't seen an improvement.

[Faa]

Well I actually tested it on my 290 and my 780ti but yea I see the point you're making

Ye exactly, when you got your gpu all the time at 99% a cpu overclock isnt going to make any difference except ocing the gpu or adding a 2nd.

 

 

Lol what? I overclocked my 3570K from 3.4Ghz to 4.8Ghz and I haven't seen an improvement.

Learn to think/read. You're sitting with a single 7950 that would make perfectly sense you won't gain any performance.

[ivhyn]

Not complety understanding why the OP chose for valley instead of unigine which is a lot more gpu bound.

Unigine is the engine...

[Faa]

Unigine is the engine...

Heaven..

[GzeroD]

How did you get it clocked that high? You must have a golden sample because my sapphire 290 max stable clock is 1108..

it was stable in the bench (same clock would give minor errors on heaven but perfectly fine in valley) at 1230 but wouldn't hold that frequency the whole time ( could drop as low as 1150 at points) i also found out when I to put a accelero xtreme iii on my brothers card that the stock cooler had a terrible tim application was horrible and when I put the stock cooler back on (he bought the wrong version) his it went from being unable to go over 1075 to hitting 1150 and throttling less.

 

Edit: also my hearing isn't amazing so I put on headphones to ignore the hurricane it produces with my fan settings. 

gaming it would hit ~76c  in a heavy bench (furmark or kombustor) ~81-85c

Another reason for base fan speed being so low is that the VRMs would get unreasonably warm on medium and low intensity games when the core temp was low and the core was over volted. 

 

 

 

[Karikiro]

Heaven..

"Unigine Corp." is the company that developed the engine for "games, simulation, visualization, serious games and virtual reality systems" called "Unigine". The two benchmarks you are referring to are called "Valley" and "Heaven" and were created by the "Unigine Corp." using the "Unigine" graphics engine.

[Faa]

"Unigine Corp" is the company that developed the engine for "games, simulation, visualization, serious games and virtual reality systems" called "Unigine". The two benchmarks you are referring to are called "Valley" and "Heaven" and were created by the "Unigine Corp." using the "Unigine" graphics engine.

And? I just forgot to add heaven in that sentence. I'm just saying that valley starts to get cpu bound with 2 high-end cards like a 780 and Heaven is nowhere to be cpu bound.

[Karikiro]

[...]

If I had a cpu that could push both to 99% there would have been a 100% fps gain.

The mainboard and memory can make a substantial difference, too. (Or even storage. Yes, I have seen systems perform better in graphical applications with SSDs instead of HDDs)

I don't think the fps would double, though.

And? I just forgot to add heaven in that sentence. I'm just saying that valley starts to get cpu bound with 2 high-end cards like a 780 and Heaven is nowhere to be cpu bound.

I just wanted to clarify. I didn't mean to offend.

 

And? I just forgot to add heaven in that sentence. I'm just saying that valley starts to get cpu bound with 2 high-end cards like a 780 and Heaven is nowhere to be cpu bound.

Yes, you might be right. I have 'tested' it and at least on two of the CPU cores the load is higher whilst running Valley compared to Heaven.

The graphics score underlines this as the Heaven score is higher than Valley's at the same presets (Extreme) Unigine provides.

Heaven

Valley

Heaven

Valley

[Faa]

The mainboard and memory can make a substantial difference, too. (Or even storage. Yes, I have seen systems perform better in graphical applications with SSDs instead of HDDs)

I don't think the fps would double, though.

Hmm in my experience an ssd made only the difference when you're getting in the game, eg like in an mmo; entering the world the ssd will allow the cpu (since it gets its info much quicker than an hdd) to push the gpu faster with information after that's done from that point there won't be any frame difference between an ssd/hdd.

Some games like Diablo I think were loading data constantly from the drive when you're playing, so yeah the minimums would be much better here. With valley/unigine when you launch it you get like 3 fps in the beginning and a few secs later you get 50 fps, this doesnt happen in every new round (like 2/18 - 3/18 - 4/18 etc).

 

 

I just wanted to clarify. I didn't mean to offend.

I know

 

 

Yes, you might be right. I have 'tested' it and at least on two of the CPU cores the load is higher whilst running Valley compared to Heaven.

The graphics score underlines this as the Heaven score is higher than Valley's at the same presets (Extreme) Unigine provides.

My cpu load was around 5-10% orsomething, but I've said it a lot; the cpu load is really irrelevant for monitoring cpu bottlenecks especially with all those multicore cpu's and the way thread&processes work it's useless to monitor. GPU's will always jump like a retard to 99% with no excuses but if it doesn't then you have a cpu bottleneck since the cpu's job is nothing more besides telling the gpu how to render a frame.

Looking at screenshots, it does exactly show the loads are higher but theyre not exactly saying that you're cpu bound. Nice analyse btw.

Anyways just to add: monitored the core loads a bit -> http://i.imgur.com/7rZguYP.jpg

[Karikiro]

[...]

My cpu load was around 5-10% orsomething, but I've said it a lot; the cpu load is really irrelevant for monitoring cpu bottlenecks especially with all those multicore cpu's and the way thread&processes work it's useless to monitor. GPU's will always jump like a retard to 99% with no excuses but if it doesn't then you have a cpu bottleneck since the cpu's job is nothing more besides telling the gpu how to render a frame.

Looking at screenshots, it does exactly show the loads are higher but theyre not exactly saying that you're cpu bound. Nice analyse btw.

Anyways just to add: monitored the core loads a bit -> http://i.imgur.com/7rZguYP.jpg

Actually, your screenshot shows fairly good that the CPU is in fact 'bottlenecking' the GPUs. I come to that conclusion, because the first core is at its limit for almost the whole time the image shows.

In the recent WAN-Show Linus mentioned that DirectX is programmed for single core operation and therefore can't utilize multiple cores at the same time.

A trick to improve performance could be to disable Intel's Hyperthreading Technology, but I am not sure about that. As far as I know, Hyperthreading allows two cores to share one cache or rather one core to split up into two seperate with one cache. If that is the case, my tip could work. I would definitely give it a try if your mainboard allows for it.

Please, let me know the results if you do try it out.

I also want to retest tomorrow. That time with the same overall framerate, assuming that then the GPU load is more even between the two benchmarks. (I have mistaken the scores of my previous test, too. The Valley score is actually higher than the Heaven score. I think this is caused by the tessellation.)