Unigine valley benchmark is getting an upgrade.

[EarthDog]

I love it when I have to search to compare.... lol

Here is my 980Ti with the 6950x @ 3.5Ghz

EDIT: I see that was with only 10t too...

Difference for those that don't want to look for it...

4.4Ghz - 4171
3.5Ghz - 4111 (also 10t)

Difference in my testing (1080p extreme 8x AA) with this setup is 1.4%.




EDIT2: Yes, I see the Precision X clock, that one may be right, not sure... you are aware that puts you likely almost 200Mhz on core clocks faster, no? What GPUz reports as boost is always VERY low, like 100Mhz low in a lot of cases. This is why I am saying to check the graph/sensor tab to see what your actual clocks were.... and why I ask in the 3DMark Fire Strike thread to report the ACTUAL boost so people have valid clocks to compare to.

EDIT3: I see this testing is at 4xAA... wonder if that makes a difference...

 

[slozomby]

ok here's 139 MHz difference

1987 vs 1848 as reported by afterburner. for around a 9 fps difference.
i7 6700k set at 4.6 no hyperthreading.

edit.
ran it with 8x aa ( still @4.6 no HT)
1987 score 4606 110.1 fps
1848 score 4589 108.2 fps
edit 2:
8xaa @4.6 w/ht.
1987 4876 score 116.5 fps
8xaa @3.5 w/ht
1987 4396 score 105.1 fps

I think it has more to do with clockspeed rather than cores or threads. I dont think my 6500 ever goes above 3.2GHz

that's what it looks like to me. but your i5 should boost to 3.6 on all cores.
make sure you bios is set to sync cores and 36 multiplier.

 

[Athlon2K15]

I think it has more to do with clockspeed rather than cores or threads. I dont think my 6500 ever goes above 3.2GHz

 

[Vayra86]

actually i found it....here.

ROFL

The 16x9 8xAA Windowed says it all really

EDIT: Also very nice to see my suspicions confirmed about running Valley at high FPS with regards to CPU clocks.

 

[EarthDog]

How was it confirmed? Considering i showed barely a difference at 100FPS and sloz showed a huge difference? How do you walk away with that conclusion?

 

[Vayra86]

How was it confirmed? Considering i showed barely a difference at 100FPS and sloz showed a huge difference? How do you walk away with that conclusion?

My point was, at high FPS the influence of CPU clocks starts to show. Precisely the difference between running the 1080 at 4xAA versus the 980ti at your 8xAA benches - sloz runs Valley at a much, much higher average FPS and his 0.9 Ghz difference in OC accounts for about 9-10% more average FPS.

 

[purecain]

holy shit my cards started working overclocked the way it should. xoc kept the power and temp linked no matter what. it gave me a big problem.

not to shabby for my second run on AB. definatly going to watercool this card with my kraken. I'm just looking at the fittings now.

 

[slozomby]

How was it confirmed? Considering i showed barely a difference at 100FPS and sloz showed a huge difference? How do you walk away with that conclusion?

easy test. I turned off AA and my gpu wouldn't hit 100% utilization but one core was pegged at 100%
when I switch to 8x aa cpu utilization stays around 60%.

 

[Zyll Goliat]

Hi guys....this is my very first post...I love to tweak my GPU´s like all of you in here ....at the moment I am stuck with "old" GTX 660TI 2GB(MSI PE OC) that I Flash bios and rise clocks and power limit.....

Current Clocks:GPU 1090Mhz/Boost 1145 Memory 1835Mhz

Here is the result:

 

[EarthDog]

easy test. I turned off AA and my gpu wouldn't hit 100% utilization but one core was pegged at 100%
when I switch to 8x aa cpu utilization stays around 60%.

View attachment 80473

Ahh, well, that is what we get for running this benchmark at 1080p and not its highest setting... a benchmark where the CPU can get in the way on only the highest of high end cards.

 

[DR4G00N]

Zotac & Gigabyte 780 Ti's in SLI @ 1250MHz/2000MHz 1.212V
Xeon X5670 @ 3.2 (Stock)

Both GPU have modded bios to remove GPU Boost & Power limit.

GPU's only at 75% usage, I need to oc the cpu to at least 4GHz again.

 

[jboydgolfer]

@GreiverBlade I thought you said that CPU speed doesn't matter much in Valley?

 

[DR4G00N]

Here's a run with 4X AA @ same settings.

GPU's only used to about 50% each lol. Only gained 0.3FPS over 8X AA.

 

[EarthDog]

@GreiverBlade I thought you said that CPU speed doesn't matter much in Valley?

It depends on the card it seems. Anything above a 980Ti (read a 1080) can see a bottleneck if there is nearly a 1000Mhz gap in modern INtel processors.

And there are people calling out Greiver?

GPU's only used to about 50% each lol. Only gained 0.3FPS over 8X AA.

You smashed into the glass ceiling because of a generations old processor and SLI. Scores would be way higher with a 66/6700K.

 

[DR4G00N]

You smashed into the glass ceiling because of a generations old processor and SLI. Scores would be way higher with a 66/6700K.

Well, no doubt but for just $100 it performs excellently when workloads can utilize all 6c/12t. (6700K is $450 here by comparison).

 

[P4-630]

It depends on the card it seems. Anything above a 980Ti (read a 1080) can see a bottleneck if there is nearly a 1000Mhz gap in modern INtel processors.

And there are people calling out Greiver?

You smashed into the glass ceiling because of a generations old processor and SLI. Scores would be way higher with a 66/6700K.

Yeah removed my previous post.
People with a new high-end card with an older CPU can see a CPU bottleneck and then overclocking can help.

 

[Vayra86]

The above results are a very good reason to base Valley scores at a more demanding setting, like the 4K setup and/or 8x AA

 

[trog100]

yep 1080 is too low for modern high end cards.. 1440 would be okay..

my machine scores nearly the same on 1440 as it does on 1080.. having said that my cpu only shows around 15% usage.. for me something aint quite right with valley..

trog

 

[slozomby]

The above results are a very good reason to base Valley scores at a more demanding setting, like the 4K setup and/or 8x AA

8x aa makes sense. i got the 1080 so i could play at 1080p with everything turned on and still have great fps. until they come out with a ultra wide x2160 monitor i don't think i'll be upgrading to 4k.

even with 8xAA theres still a cpu limit when i downclock 1ghz (to replicate the i5-6500) . but its a much smaller % difference .

yep 1080 is too low for modern high end cards.. 1440 would be okay..

my machine scores nearly the same on 1440 as it does on 1080.. having said that my cpu only shows around 15% usage.. for me something aint quite right with valley..

trog

I'm not sure how you're accomplishing that. with 4xaa i drop from 145fps at 1080p to 87fps at 1440p. unless dsr is adding stupid amounts of overhead

 

[trog100]

i am running sli.. and as i say something aint quite right with valley on my machine.. what i aint entirely sure.. i think my 1440 score is about right and 1080 score way too low..

trog

 

[P4-630]

(to replicate the i5-6500)

Well I own a i5 6500 and this was my score:

 

[Athlon2K15]

Not putting anyone down here, but the score list is formatted something terrible. Doesn't someone know bbcode,php or HTML that we can inject a nice looking table? And yes we really need to separate resolutions as 1080p is weak and cant stress a 1080 worth a damn. Could we have three tables with 1080p, 1440p and 4K scores?

 

[slozomby]

Well I own a i5 6500 and this was my score:

View attachment 80512

almost exactly what i got with my 1080 when i downclocked it.

i am running sli.. and as i say something aint quite right with valley on my machine.. what i aint entirely sure.. i think my 1440 score is about right and 1080 score way too low..

trog

i didn't see your scores. what were they?

 

[P4-630]

almost exactly what i got with my 1080 when i downclocked it.

View attachment 80513

I'm curious... To what clocks?
Do you still have the full original screenshot?
On that I can see your card's clocks and temp.

 

[slozomby]

I'm curious... To what clocks?
Do you still have the full original screenshot?
On that I can see your card's clocks and temp.

sorry i meant when i downclocked the cpu to 3.2 w/o HT. the 1080 stayed at the same clocks i normally run 1987 MHz
temps on the card is always ~60c give or take a coupla degrees while folding for days on end

here's afterburner profile 1