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

[Face2Face]

I don't suppose there would be any way to, ah, well, add another column for the GPU voltage? [runs and hides...]

:biggrin:

 

[Elfear]

I changed the results layout. I added a ranking column, let me know if you guys like this better or of you like the old one.?

Looks good. Thanks for putting that together.

 

[nico9]

There's mine for a single HD 7970 using the 13.10 Beta driver.

1260|1661 = 54.5 FPS



Direct link-
http://img46.imageshack.us/img46/369/w4nm.jpg

 

[imaheadcase]

Behold the power! Don't PM me for offers, this baby is not for sale!

1xGTX 580 772mhz 1002 mem 1544 shader

 

[Elfear]

There's mine for a single HD 7970 using the 13.10 Beta driver.

1260|1661 = 54.5 FPS



Direct link-
http://img46.imageshack.us/img46/369/w4nm.jpg

That's a seriously high score for the clocks you ran. Just a tick under Durvelle's run at 1325/1850.

 

[Durvelle27]

That's a seriously high score for the clocks you ran. Just a tick under Durvelle's run at 1325/1850.

somethings not right O.O

 

[Face2Face]

SNIP

Please resubmit without tweaks.

 

[nico9]

Please resubmit without tweaks.

This time I read the OP. I didn't realize you restricted use of the CCC. I was using performance driver settings on the previous run straight from the CCC and I warmed the bench by cycling through all stages with enter before I benched.

Do you allow people to set +20% under power control in the driver? That's not default and if I don't set +20% my driver may throttle my 7970 when it's OCed.

 

[(sic)Klown12]

Here's what my old-in-the-tooth 7970 with a mild OC can deliver.

 

[Face2Face]

This time I read the OP. I didn't realize you restricted use of the CCC. I was using performance driver settings on the previous run straight from the CCC and I warmed the bench by cycling through all stages with enter before I benched.

Do you allow people to set +20% under power control in the driver? That's not default and if I don't set +20% my driver may throttle my 7970 when it's OCed.

OC your card as much as you can Any % power limit is allowed.

 

[Grooveriding]

Got around to doing a run with both my cards installed. My replacement card is not the best clocker. Can only manage 1300 on 1.3V and seems to not be able to even do 1350 at 1.35V. Maybe could of done a bit more, but I got lazy and stuck to what I knew worked.

780 SLI
1410/1800
1306/1800

I noticed there is a score a smidge higher than mine with 780s at 1228/1800.. what sort of voodoo is going on with that score ?

 

[YBS1]

Got around to doing a run with both my cards installed. My replacement card is not the best clocker. Can only manage 1300 on 1.3V and seems to not be able to even do 1350 at 1.35V. Maybe could of done a bit more, but I got lazy and stuck to what I knew worked.

780 SLI
1410/1800
1306/1800

I noticed there is a score a smidge higher than mine with 780s at 1228/1800.. what sort of voodoo is going on with that score ?

Nice score! I'll try to knock you off once my new psu arrives, finally shipped out today. Sucks about your replacement card though. Your cards are under water though aren't they?

 

[Teizo]

I noticed there is a score a smidge higher than mine with 780s at 1228/1800.. what sort of voodoo is going on with that score ?

Yeah, that is definitely off. Maybe he ran it with the NV control panel set to high performance or something before the no tweaking was implemented.

 

[Teizo]

My memory ended up topping out at 7600. I can't get the core over 1293 unfortunately, even with the board power all the way up. I'd have to mod the bios to allow for 1.25 volts, or higher I imagine, to get over 1300. Not sure I really want to do that since these cards only have one bios each.

88.4/3699 with clocks of 1293-1280/1900(7600). Cpu @ 4.4ghz...Memory @ 2133mhz 9-1T.

 

[Face2Face]

Updated

 

[omeds]

Please put my score back, it is 100% legit. I have one of if not the biggest (stable) video memory oc and the fastest CPU in the chart, which is why I have edged out several others.

 

[Tempered81]

post pics of your CPU oc

 

[Tempered81]

omeds, looks like they removed your score because it doesnt jive with the submissions from the rest of us. your 4.8 quad core with 1200/1850 780s doesnt score higher than a 5g 6 core with 1410/1850 780s. So you have to share your hax & tweaks. This bench is so easy to hack anyway.

 

[omeds]

My vid memory oc was at +680mhz, so ~7.38GHz I suppose. I wouldn't be surprised if other 780 scores were with unstable vid mem oc and error checking was kicking in.

Secondly, 4770k at 4.8 is faster than any current 6 core chip at 5GHz in Valley (or any app that uses 4 cores or less) so my peak fps are probably higher.

CPU oc as requested:

I am using 1.366v and 2666mhz cas 10 RAM now though.

 

[Face2Face]

Please put my score back, it is 100% legit. I have one of if not the biggest (stable) video memory oc and the fastest CPU in the chart, which is why I have edged out several others.

Hey Omeds,

I asked another member to clock their GTX 780's in SLI at the same clock speeds you are running. I understand that your CPU is faster clock per clock, but I am having a hard time understanding the 10 FPS delta. 10 FPS in this benchmark is a huge amount to gain just from a CPU alone. Looks at Durvelle's 7970 with his 8350 and Elfears 7970 with his 4670K - Even at different GPU clocks speeds and different CPU architectures the scores are pretty close. Like I said, you must have something unknowingly selected in your Nvidia control panel to make your score higher than your competition. If anyone else can make sense of this then you are welcome to chime in. And if another member running the same gpu setup can post pictures of their nvidia control panel that may be helpful as well.

At Face2Face's request, here is what a 780SLI@1228/1800 score should look like (give or take a few points), keep in mind I'm running this bench on a 3930K@5.0Ghz and Valley does scale at least somewhat with cpu speed unlike Heaven.

 

[omeds]

The single GPU benchmarks you mentioned would not be CPU bound in any areas, but with CFX/SLI there will be more instances of hitting CPU bound parts of the bench (the points where there is least load on the GPU's and fps are very high).

I also think some of the 780 scores are not in fact using 100% stable memory oc's. For example if I clock to 7.4GHz my score goes down 10-20% from error checking, where as 7.38GHz seems to produce my highest scores. Just because you can pass Valley with uber mem oc doesn't mean it's stable or giving you the right score.

 

[Face2Face]

The single GPU benchmarks you mentioned would not be CPU bound in any areas, but with CFX/SLI there will be more instances of hitting CPU bound parts of the bench (the points where there is least load on the GPU's and fps are very high).

I also think some of the 780 scores are not in fact using 100% stable memory oc's. For example if I clock to 7.4GHz my score goes down 10-20% from error checking, where as 7.38GHz seems to produce my highest scores. Just because you can pass Valley with uber mem oc doesn't mean it's stable or giving you the right score.

Interesting... Have you ran the benchmark with lower CPU clocks? If so, does it make that much of a difference?

 

[omeds]

I haven't run at lower CPU clocks, I'll give it a shot and see if there are any changes.

 

[omeds]

At 4.2 I scored 130.3 with a max of 213 (vs 137 and max of 235 at 4.8)

 

[Face2Face]

At 4.2 I scored 130.3 with a max of 213 (vs 137 and max of 235 at 4.8)

That is a huge jump I wasn't expecting that, but again you are running two cards. One more thing, can you make sure this is set to let 3D application decide..

http://www.tweakguides.com/NVFORCE_6.html