• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

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

Page 16 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
guskline said:
I'm not going to do a "Kamakazi" Valley run just to beat a score.
Click to expand...

I always test for stability with Kombuster before I run any benchmark. If it isn't stable and rendering properly, like others mention...no point whatsoever to run the benchmark. Benchmarks aren't a good tool to test for stability.
 
Teizo said:
I always test for stability with Kombuster before I run any benchmark. If it isn't stable and rendering properly, like others mention...no point whatsoever to run the benchmark. Benchmarks aren't a good tool to test for stability.
Click to expand...
Excellent suggestion.
 
Teizo said:
I always test for stability with Kombuster before I run any benchmark. If it isn't stable and rendering properly, like others mention...no point whatsoever to run the benchmark. Benchmarks aren't a good tool to test for stability.
Click to expand...

Interesting. My 7950 will run Kombuster at 1300 mhz but will not run valley at that speed.
 
Teizo said:
I always test for stability with Kombuster before I run any benchmark. If it isn't stable and rendering properly, like others mention...no point whatsoever to run the benchmark. Benchmarks aren't a good tool to test for stability.
Click to expand...

I don't think Furmark/Kombuster are a good way to test for stability at all. They stress your cards shaders more than anything else and put an unnecessary load on your cards vrms. Pretty sure most on here don't use either one to test stability.

In my experience to test OC stability is to run a multitude of benchmarks (3Dmark/Unigine) and games. I find demanding games like Crysis 3 to stress my cards and system to max . It will max both of my 7950's out and keep my 3570 @ 80-90% If it fails in Crysis 3, then to me it's not stable.
 
Face2Face said:
I don't think Furmark/Kombuster are a good way to test for stability at all. They stress your cards shaders more than anything else and put an unnecessary load on your cards vrms. Pretty sure most on here don't use either one to test stability.

In my experience to test OC stability is to run a multitude of benchmarks (3Dmark/Unigine) and games. I find demanding games like Crysis 3 to stress my cards and system to max . It will max both of my 7950's out and keep my 3570 @ 80-90% If it fails in Crysis 3, then to me it's not stable.
Click to expand...

Well that is interesting. I've never had Unigine fail if I ran Kombuster first and allowed enough time for any lock ups, etc. to set in. I used to run 3D Mark a long time ago, but Valley does stress the cards well so that is pretty much the only one I run (although I have run Heaven several times because I like it as well).

I guess what I meant to say was that making one pass of Valley is not a sign of stability. You need to let it loop several times. I don't have Crysis 3 since I've been out of the loop for a while (only playing Blacklist atm), but it is on deck because I really enjoyed the first one.
 
Teizo said:
I guess what I meant to say was that making one pass of Valley is not a sign of stability. You need to let it loop several times. I don't have Crysis 3 since I've been out of the loop for a while (only playing Blacklist atm), but it is on deck because I really enjoyed the first one.
Click to expand...

Agreed on the looping. You should look on [H] and here to see if you can get a copy for $10 or $15. It's the most beautiful game out there IMO. It will stress your setup to max. The gameplay is not as good as Crysis 1 though 🙁
 
Well, unless I can get my memory any higher...this is probably going to be about as high as I can get. I upped the core 5mhz, but more importantly upped the vcore +12 which allowed a higher boost...the 1293/1280 I was shooting for. It is interesting to see my 2nd, and less performing, of the two cards only had a TDP of 82% compared to 120.6% from the primary GPU. I never paid attention to that before, but I guess that is why switching the cards around helped my clocks.

Was only .5 fps from what I was shooting for. Maybe I can get my RAM higher a little later and squeeze out that last little bit as I am learning just what this new card is capable of...For the time being though, I imagine this will be it.

3514 with clocks of 1293-1280/(1675)6700

-edit

I was able to get my memory up to 6800mhz after all...so I think I'm done now. Any higher and I'm just going to be using up a whole lot more wattage for very little return.

3549 with clocks of 1293-1280/1701(6800)

UnigineValley1293-1280_6800_zps004a709c.jpg~original


Thanks for the tip on Crysis 3 Face2Face. I was prepared to pay $40, but if I can get a key for less that will be great. I look forward to playing it maxed out with this SLI set up.
 
Last edited:
Teizo said:
SNIP
Click to expand...

Your welcome, though playing it maxed out @ 1080P is probably not gonna happen 🙁 Very high settings with 2XSMAA is realistic with your setup, the game is very demanding. Can you ramp up your fan speeds and raise your temp limit to get a higher core clock to beat the uber clocked GTX 780?
 
Face2Face said:
Your welcome, though playing it maxed out @ 1080P is probably not gonna happen 🙁 Very high settings with 2XSMAA is realistic with your setup, the game is very demanding. Can you ramp up your fan speeds and raise your temp limit to get a higher core clock to beat the uber clocked GTX 780?
Click to expand...

Seems like no matter what I do, once I get to 1306, they lock up. It may be a limitation of this pcb? I don't know. The 760 Hawks have a better pcb and power delivery and a review I saw they were boosting to 1365 at the regular voltage of 1.2. I don't have any heat issues since the top card only gets to 71F or so with the fan speed @ 70% to 80%. I have kept my temp limit @ 80F though, so maybe raising it will help.

-edit

However, as it turns out...I was able to get my memory up to 6800mhz so I was able to get 84.8fps and get just above what I was shooting for. Ran through three times in a row. Not sure I'm entirely stable, but it's good enough for the purposes of this thread I reckon.

See updated pic in previous post.
 
Last edited:
Was actually able to get my memory up to 7.3ghz on this new card. 7.5ghz artifacted Kombuster, and 7.4ghz crashed Valley. 7.3ghz ran just fine though. I know I said I was done in the previous post lol, but decided to push for just a bit more anyway. Surprised it went so high. Pretty sure I have reached the peak of this set up now.

3612 with clocks of 1293-1280/1825(7300)

Unigine1293-1280_7300_zps97cf9056.jpg~original
 
No tweaks allowed 🙁? Aww... Well you didnt specify in OP so i assumed it was allowed, and yea i used OCns tweaks. Nothing else.
 
Why do people buy 760s? As if those cards arent [redacted] as is but will be even more [redacted] now that basically all games will be optimized for AMD due to consoles using AMD... Just look at BF4; 780 is the same as a 7970 in that game...

Ugh people these days...

Warning issued for inappropriate language.
-- stahlhart
 
Last edited by a moderator:
Rangerjr1 said:
Why do people buy 760s? As if those cards arent [redacted] as is but will be even more [redacted] now that basically all games will be optimized for AMD due to consoles using AMD... Just look at BF4; 780 is the same as a 7970 in that game...

Ugh people these days...
Click to expand...


Well, aren't you a lively one.
 
Last edited by a moderator:
Rangerjr1 said:
Ok, so what should driver settings be? Tessellation use application settings or AMD optimized?
Click to expand...

I started this thread before the tweaking was really catching on. Most of the scores submitted are without any tweaks. Basically install Valley, OC your GPU and run it. So I assume the settings in the control panel are going to be set to "use application settings" Just trying to make this as fair for comparison sake. Thanks RangerJr1, you shouldn't worry too much, pretty sure you still have the fastest 7970 on here 😉

Rangerjr1 said:
Why do people buy 760s? As if those cards arent [redacted] as is but will be even more [redacted] now that basically all games will be optimized for AMD due to consoles using AMD... Just look at BF4; 780 is the same as a 7970 in that game...

Ugh people these days...
Click to expand...

Let's play nice Ranger 🙂 The mods are not going to be kind to this....
 
Last edited by a moderator:
Face2Face said:
I started this thread before the tweaking was really catching on. Most of the scores submitted are without any tweaks. Basically install Valley, OC your GPU and run it. So I assume the settings in the control panel are going to be set to "use application settings" Just trying to make this as fair for comparison sake. Thanks RangerJr1, you shouldn't worry too much, pretty sure you still have the fastest 7970 on here 😉
Click to expand...

Yea i understand you want it to be fair, and adding tweaks now would be too much work i guess.

What do you want me to do? Run a quick run on valley with default driver settings and set all i can to "use application settings" ?
 
Rangerjr1 said:
Yea i understand you want it to be fair, and adding tweaks now would be too much work i guess.

What do you want me to do? Run a quick run on valley with default driver settings and set all i can to "use application settings" ?
Click to expand...

That would be great.
 
Rangerjr1 said:
Why do people buy 760s? As if those cards arent [redacted] as is but will be even more [redacted] now that basically all games will be optimized for AMD due to consoles using AMD... Just look at BF4; 780 is the same as a 7970 in that game...

Ugh people these days...
Click to expand...

:colbert:
 
Last edited by a moderator:
5ken3a.jpg


otnkbb.jpg


Lower clocks and no tweaks, still top 7970 😀

Also, by these results. I cant help but think that the 770 guy has tweaked his score, a 770 with lower clocks AND lower IPC cant score higher.
 
Last edited:
Rangerjr1 said:
Lower clocks and no tweaks, still top 7970 😀

Also, by these results. I cant help but think that the 770 guy has tweaked his score, a 770 with lower clocks AND lower IPC cant score higher.
Click to expand...

Looks good :thumbsup:. I can PM John, the guy with the GTX 770 and see. I am pretty sure he wasn't tweaking, as his FPS can probably be attributed to the High memory frequency that Valley seems to love. I will check though, thanks for keeping it real Ranger.
 
Face2Face said:
Looks good :thumbsup:. I can PM John, the guy with the GTX 770 and see. I am pretty sure he wasn't tweaking, as his FPS can probably be attributed to the High memory frequency that Valley seems to love. I will check though, thanks for keeping it real Ranger.
Click to expand...

Still does not make sense, because even at high frequency the bandwidth is a lot lower than a small OC on the RAM of a 7970 because of the bus width (7970 has 384 and GK104s has 256) GK104s max out at around 250GB/s while 7970 around 360GB/s... Something is not right.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top