Unigine Valley 1.0 Benchmark Thread ** Post your scores**

[Saffron]

I ran two benchmarks, one at all stock clocks and one with a CPU+GPU overlock. I just wanted add a little more info for comparisons. I will post my MSI Notebook with an i7-4700MQ + GTX770M later on.
CPU: i5-3570K
GPU: MSI Lightning GTX680 2GB

Here's my results:

#1 Stock Clocks

#2 Overlocked: CPU@4.6GHz and GPU@1369/3104

 

[justin4pack]

Stock clocks.

Overclock: CPU @ 4.5GHZ GPU @ 1150/3055

 

[Belkov]

On stock:

 

[Yungbenny911]

Please confirm you are NOT using tweaks. Thanks!

Scores updated.

Oh shoot i'm sorry, i didn't read it OP lol. Yes those are with tweaks.

 

[Tristor]

My results:

Completely unmodified GTX 780 Classified w/ ACX Cooler in SLI

According to the specs it's 993 Core, 1046 Boost, 6008 Mem. According to my Afterburner logs though, it was 1127 and 1134 on GPU1 and GPU2 respectively for their actual maximum core clocks during that run. I'm guessing they ramped up early before temps rose and then became thermally limited, especially GPU1 which is the top card and has restricted airflow due to the minimal spacing on my motherboard (soon to be rectified).

 

[rtsurfer]

My results:

Completely unmodified GTX 780 Classified w/ ACX Cooler in SLI

According to the specs it's 993 Core, 1046 Boost, 6008 Mem. According to my Afterburner logs though, it was 1127 and 1134 on GPU1 and GPU2 respectively for their actual maximum core clocks during that run. I'm guessing they ramped up early before temps rose and then became thermally limited, especially GPU1 which is the top card and has restricted airflow due to the minimal spacing on my motherboard (soon to be rectified).

I am interested in what ASIC you got & also max individual OC of each card.

 

[Face2Face]

Face2Face, do you want me to create a Doc with your results and make graphs with them.???

Thanks KaRLiTos, but I am good for now. Probably just going to keep this thread simple for the time being.

Oh shoot i'm sorry, i didn't read it OP lol. Yes those are with tweaks.

No worries, if you want to edit your scores without tweaks just let me know.

 

[Tristor]

I am interested in what ASIC you got & also max individual OC of each card.

The top card reads 75.3%, and the bottom card reads 71.0% in GPU-Z 0.7.7 for ASIC quality. They both have Hynix memory (not Samsung, but at least it's not Elpida). I haven't done any overclocking of the GPUs yet because I'm still trying to figure out exactly what my next step is in order to resolve my particular dilemma.

 

[Yungbenny911]

No worries, if you want to edit your scores without tweaks just let me know.

This was a quick run without any tweaks what so ever :thumbsup:

x2 GTX 770 @ 1423Mhz (core), 2004Mhz (mem)

 

[CropDuster]

What is the significance of the ASIC score in GPUZ? My 290 is 82.2%

 

[Phallacy]

Hey everyone this is my unigine valley run. Second run I had with the benchmark. Was only using heaven prior.

290x X3 watercooled
1. Sapphire Tri X OC 290x 1215/1500 hynix
2. Sapphire 290x ref 1150/1350 elpida
3. XFX 290x 1250/1550 hynix

 

[rtsurfer]

Alright so I am finally done testing,

My username is rt123 @ OCN, so you know Face2Face

Core i7 4770K @ 4.6 Ghz

EVGA GTX 780 Classified, ASIC 66.9%, Memory Hynix

So continuing our discussion about Samsung vs Hynix Memory

I got 75.5 Fps in Valley on my now deceased/return 780 FTW with Samsung memory at 1241/1824

Here is my Classy at the same clocks

Almost the same, but then I remembered that my FTW was throttling and going back & forth between 1228 - 1241, so here is the Classy at 1228 / 1824

0.2 FPS is within margin of error, so I think we can conclude that on the 780s there is no difference between Hynix & Samsung memory at the same clockspeed.

Hynix does not offer lower performance due to some loose timings facade.

Here is a demonstration of how well valley scales with memory

1241 / 1927

Finally here is my max benchmarks stable OC at room temperature

Using the classy tool

Core = 1.25 V
Memory = 1.64 V

Frequency = 1306 / 1927

I can clock further with more voltage but rather not risk it on air.
You can remove my 780 FTW score if you want to as the card has been returned.

 

[Face2Face]

SNIP

Hey RTsurfer, thanks for doing the comparison. It's nice to see that Hynix vs. Samsung at the same clock speeds are the same. However, I never argued that point and never jumped on the memory timing bandwagon. What I was saying is that on "avg" Samsung memory will clock higher than both Hynix and Elpida. I don't know if Hynix is using a new sku for the GTX 780ti's because their memory requirements are much higher out of the gate? And if they are now using a different sku throughout the GK110 line? But it does seem that people are having better luck obtaining higher clock speeds with their Hynix cards now than I have seen in the past.

 

[rtsurfer]

Hey RTsurfer, thanks for doing the comparison. It's nice to see that Hynix vs. Samsung at the same clock speeds are the same. However, I never argued that point and never jumped on the memory timing bandwagon. What I was saying is that on "avg" Samsung memory will clock higher than both Hynix and Elpida. I don't know if Hynix is using a new sku for the GTX 780ti's because their memory requirements are much higher out of the gate? And if they are now using a different sku throughout the GK110 line? But it does seem that people are having better luck obtaining higher clock speeds with their Hynix cards now than I have seen in the past.

Don't worry I wasn't pointing at you for the "looser timings" facade.
It seems to be the general conscensus at OCN, that is what I was pointing at.

I have question for you though.

Have you hard modded your lightning..??

From what I read there is no way to go above 1.25 V on a 780 lightning without hard modding.

 

[Face2Face]

Don't worry I wasn't pointing at you for the "looser timings" facade.
It seems to be the general conscensus at OCN, that is what I was pointing at.

I have question for you though.

Have you hard modded your lightning..??

From what I read there is no way to go above 1.25 V on a 780 lightning without hard modding.

No worries - You should post your findings on OCN, maybe clear up some of the FUD.

With the Lighting, if you want to get over 1.2v you need use the Rbby Tool. More of a softmod. I have taken my card up to 1.485v with my wanna-be water cooling setup.. around 1.465v with vdroop. Pretty sure this card would take 1.5v+ like a champ, but I would need a full cover waterblock for me to do so, which should net me in the 1550-1600MHz range.

http://www.overclock.net/t/1398725/...s-on-lots-of-cards-some-with-llc-ab-b-18/0_50

 

[rtsurfer]

No worries - You should post your findings on OCN, maybe clear up some of the FUD.

With the Lighting, if you want to get over 1.2v you need use the Rbby Tool. More of a softmod. I have taken my card up to 1.485v with my wanna-be water cooling setup.. around 1.465v with vdroop. Pretty sure this card would take 1.5v+ like a champ, but I would need a full cover waterblock for me to do so, which should net me in the 1550-1600MHz range.

http://www.overclock.net/t/1398725/...s-on-lots-of-cards-some-with-llc-ab-b-18/0_50

About posting on OCN, I 'll either link this post or make a separate post next time someone mentions "Hynix loose timings". Too lazy do it now.


And thanks for the lightning voltage explanation.:thumbsup::thumbsup::thumbsup:

You should start saving up for a Custom loop. There is where I am headed.

 

[Face2Face]

About posting on OCN, I 'll either link this post or make a separate post next time someone mentions "Hynix loose timings". Too lazy do it now.


And thanks for the lightning voltage explanation.:thumbsup::thumbsup::thumbsup:

You should start saving up for a Custom loop. There is where I am headed.

Good idea and no problem. I would, but since I am stuck with Eplida on my card and can't get over 1650MHz on the memory, there really isn't much point... Looking forward to Maxwell though. Seeing how well the GTX 750ti overclocks has me excited for a flagship Maxwell - Classified or Lightning. Maybe AMD will offer something similar clocking? Loved my HD 7950 :wub:

 

[Belkov]

What's up Face? Long time no see...

This one is for you:

1150/1550

780s are way better in the unigine benches...

 

[Face2Face]

What's up Face? Long time no see...

This one is for you:

780s are way better in the unigine benches...

Hey Belkov, hope you're doing well. Nice card you got there :thumbsup: You may wanna upload the img again, for some reason it's not coming through.

 

[Belkov]

Can you see it now, mate?

Loved my HD 7950 :wub:

LOL, I missed my 7950 too... Poor little thing - it is minnig a lot these days. I sold it to a friend who mines 24/7 with it...:'(

 

[Face2Face]

Can you see it now, mate?

LOL, I missed my 7950 too... Poor little thing - it is minnig a lot these days. I sold it to a friend who mines 24/7 with it...:'(

Still nothing..

I'm sure my old 7950 is mining now as well.. Sucks though, because I sold it a month before the mining explosion happened... I could of sold it for more than I paid for it :\

 

[Belkov]

Yes, i could sold it for more too, but somehow i am not such a person...
Still i took good money for it.

Well, i will try to upload the screenshot somewhere else.

Edit: How is it now? I uploaded it in Dropbox, but i only can share links.

 

[Face2Face]

Yes, i could sold it for more too, but somehow i am not such a person...
Still i took good money for it.

Well, i will try to upload the screenshot somewhere else.

Edit: How is it now? I uploaded it in Dropbox, but i only can share links.

Got it. Thanks, I will update the results sometime in the next day or so.

 

[MentalIlness]

 

[Belkov]

Got it. Thanks, I will update the results sometime in the next day or so.

Oh, yes, this score is with no tweaks with the drivers. With tweaks i can achieve 4-5 fps more.
The above is due to:

Please confirm you are NOT using tweaks. Thanks!