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

[Najuno]

With electricity, every material has a certain level of resistance, and that resistance needs to be overcome by adding more force (or voltage) behind the charge of electrons (amps). the ASIC quality tells the bios of the graphic card how much voltage the card needs compared to the theoretical design. the VRM's regulate the voltage so that the gpu gets the voltage it needs based on its ASIC score, instead of every card getting the same voltage indiscriminately. In theory a higher ASIC score means better for air cooling, because there is less resistance, and less heat generated as a result (resistance causes heat.. electric heaters work in this manner). Temperatures also effect the resistance, so switching to water cooling automatically lowers the resistance of the gpu, and less voltage is needed to maintain stability. 

 

It's more complicated than this but on a base level it makes sense. ASIC% should not effect overclockability.

The ASIC % has besides your power connectors, quality of VRMs and memory controller the biggest influence on the OCability of a GPU. The ASIC determines what the GPU is capable to do in terms of OCing or not OCing and running really efficient with base clock and even lower voltage than out of the box.

 

ASIC between 100-95% are super efficient but not the best for OCing.

ASIC between 95-80% are usually good OC samples for air

ASIC between 79-65% are optimal for using water cooling

and last but not least best samples for LN2 are samples under 65%.

 

The lower the ASIC is the more power you need to overcome the electrical resistance. In the first place that sounds bad and without sufficient cooling it is. But in terms of really high OCing it's good because the more voltage and watts you apply the more brutal you can go in terms of OCing. Od course this isn't as easy as I try to explain this, because this depends on a lot of factors. Like expensive cooling solutions, BIOS modding, power delivery from the PSU and so on.

 

EDIT: This doesn't really matter to most people out there, I don't know too many that use modded BIOS on their GPUs or that are even skilled enough and capable of modding BIOS on their own. So most people buy a GPU use what they've got with MSI Afterburner or EVGA Precision X and call it a day when they've got a stable result. In reality most of their GPUs could probably OC higher.

[Raudi_]

Is that the kepler version of the 630? if it is I'm disappointed, I thought that would beat my 545.

Yeah its a GT 630 i would never beat the 545. since the GT line of cards are like cheapy mc cheap cheap cards.

[Lord Pantaloons]

Yeah its a GT 630 i would never beat the 545. since the GT line of cards are like cheapy mc cheap cheap cards.

there are three versions of the GT 630, do you know which version it is? http://www.geforce.com/hardware/desktop-gpus/geforce-gt-630/specifications

[Raudi_]

there are three versions of the GT 630, do you know which version it is? http://www.geforce.com/hardware/desktop-gpus/geforce-gt-630/specifications

GT 630 G5 is the model. of this one.

[Lord Pantaloons]

GT 630 G5 is the model. of this one.

ah, I bet the kepler versions faster than my 545 due to the 384 cuda cores compared to the 96 in the other versions.

[Raudi_]

ah, I bet the kepler versions faster than my 545 due to the 384 cuda cores compared to the 96 in the other versions.

Yeah I am looking at getting my hands on the kepler version and then on some more kepler and fermi cards for benchmarking.

[pdv789]

ah okay, that clears it up alot for me. Thank for the explanation, sorry I couldn't reply earlier started reading articles behind asic and fell asleep XP

[Briggsy]

The ASIC % has besides your power connectors, quality of VRMs and memory controller the biggest influence on the OCability of a GPU. The ASIC determines what the GPU is capable to do in terms of OCing or not OCing and running really efficient with base clock and even lower voltage than out of the box.

ASIC between 100-95% are super efficient but not the best for OCing.

ASIC between 95-80% are usually good OC samples for air

ASIC between 79-65% are optimal for using water cooling

and last but not least best samples for LN2 are samples under 65%.

The lower the ASIC is the more power you need to overcome the electrical resistance. In the first place that sounds bad and without sufficient cooling it is. But in terms of really high OCing it's good because the more voltage and watts you apply the more brutal you can go in terms of OCing. Od course this isn't as easy as I try to explain this, because this depends on a lot of factors. Like expensive cooling solutions, BIOS modding, power delivery from the PSU and so on.

EDIT: This doesn't really matter to most people out there, I don't know too many that use modded BIOS on their GPUs or that are even skilled enough and capable of modding BIOS on their own. So most people buy a GPU use what they've got with MSI Afterburner or EVGA Precision X and call it a day when they've got a stable result. In reality most of their GPUs could probably OC higher.

Yeah. In the case of a high asic score it requires tricky bios modding to get voltages up. the asic value is stored in the bios, and by lowering the asic value stored in bios (via flashing the bios) the actual voltage delivered to the core will be increased. Up until you reach a voltage hardlock (ie nvidia).

Anyone doing this does so at their own risk.

[Cheddle]

I think its a bit odd though that a card designed and binned for overclocking is still hard locked, I'm hoping someone will find a way to do a voltage edit like with 7xx series cards for the 970/980

 

 

 I agree for sure.

 

I think it may have to do somehow with how the vrms are designed, because I think every 970 right now has the exact same VRM power design, but some have "better" quality components.

 

I think all of them are 6+2 or something like that.    I was hoping EVGA or someone would make a 970 that's binned + has a overbuilt VRM, but nobody's done it yet which makes me sad.

 

Nvidia green light all AIB designs, their GPUboost 2.0 is a component of these designs and it dictates how much voltage can be added and at what temperatures.  Custom BIOS sorts the voltage limits... Nvidia are in my opinion very conservative with their voltage/temp/clock tables and 3rd party manufacturers are surely held back by their rules.

 

I flashed my 980 and since I can run a further +25mv on the core, and my voltages don't throttle as early in the temp range. I am sure gigabyte would have liked to have done this from the get go...

[Lays]

Nvidia green light all AIB designs, their GPUboost 2.0 is a component of these designs and it dictates how much voltage can be added and at what temperatures.  Custom BIOS sorts the voltage limits... Nvidia are in my opinion very conservative with their voltage/temp/clock tables and 3rd party manufacturers are surely held back by their rules.

 

I flashed my 980 and since I can run a further +25mv on the core, and my voltages don't throttle as early in the temp range. I am sure gigabyte would have liked to have done this from the get go...

 

 

I'm not sure what you mean by custom BIOS sorts the voltage limits, we already discussed what was going on with the voltage lock being hardware related & not BIOS related.

[Cheddle]

In other news I got "gifted" a core2 duo powered laptop that I intend on turning into a mineral oil build - has an unknown GPU but I will definitely be benchmarking it! 

 

I want your opinion: What should I put under mineral oil

 

Should I do the laptop - core 2 duo, 2gb ddr2, unknown GPU

 

or

 

my spare parts - AMD Athlon 64 X2 3800+, 4gb ddr2, and either an ATI 4850, ATI 5450, Nvidia 8600 GS or my random fake ass chinese crapcard (9500 GS?)

 

The PC will be used for nothing. Maybe a guest PC if someone wants to use the internets. a benifti would be to be able to play LoL - so the 4850 would be the go...

 

If this goes well I will put one of my 1u rack servers under oil, maybe make a table out of it or a wall feature

[Cheddle]

I'm not sure what you mean by custom BIOS sorts the voltage limits, we already discussed what was going on with the voltage lock being hardware related & not BIOS related.

 

My comment was directed at some 980's not being able to reach their maximum core voltage due to BIOS limits. I only skimmed over the prior posts. My card is soft limited. not sure if the 970 G1 gaming suffered the same flaw.

[pdv789]

Let me redo my overclock now that I have gpu-z, I had previously only been using afterburner. Gpu-z has a nice feature that tells me how much of the tdp the card is pulling       Afterburner only shows 0 for no and 1 for max tdp .... not very helpfull there lol

[pdv789]

It seems my driver stability is causing my card to crash. I do anything above +200 on the core and it instantly crashes as soon as I launch any program even though I have 100% stability at 200. I'm thinking if I go with the non qhl driver I might get better stability or just doing a wipe on my pc just to get a higher overclock. I'm entertaining the thought of bios hacking though, I only reach 82% tdp at any given time according to gpu-z with my max overclock. I'm not sure how accurate it is but it seems reliable as I only do 65% tdp until I raise the voltage and overclock.

[Jumper118]

280x crossfire at stock

fx8350 at 4.5 ghz

preset is extreme hd

 

 

 

 

need a gpuz screenshot please

 

 

 

Thread updated at last!

[WaitForMeImOnDialUp]

need a gpuz screenshot please

 

 

 

Thread updated at last!

[Jumper118]

 

thanks

[Raudi_]

thanks

I think I will stay the lowest scoring single card with 4GB of VRAM for 1440p

[Raudi_]

Basic:


Extreme HD(Was the hardest because of exposed memory on the back of the card):

[Lord Pantaloons]

-bloop-

There was someone on the forum a little while ago that posted about his Ti boost being the fastest he's seen but it isn't anymore. Theirs was in the 1200Mhz range I think but I can't quite remember who it was.

[Raudi_]

There was someone on the forum a little while ago that posted about his Ti boost being the fastest he's seen but it isn't anymore. Theirs was in the 1200Mhz range I think but I can't quite remember who it was.

Yeah if I had a MSI card I would be even faster then what I got with this DCUII card.

[NannerBeans]

Intel Core i7 4770k

GTX 780 ti

Core Clock: 1310.7 MHz

Memory Clock: 1749.6MHz 

Avg FPS: 79.5

Score: 3328

Setting: Extreme HD

 

 

[MourningSouls]

Did I do this right?

 

Intel i5-4690
EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz
74.5
1877
Extreme

 

http://i.imgur.com/RgrbFnK.jpg

 

Intel i5-4690
EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz
217.7
5485
Basic

 

http://i.imgur.com/A5GwTaR.jpg

[Briggsy]

Did I do this right?

 

Intel i5-4690

EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz

74.5

1877

Extreme

 

http://i.imgur.com/RgrbFnK.jpg

 

Intel i5-4690

EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz

217.7

5485

Basic

 

http://i.imgur.com/A5GwTaR.jpg

 

when posting you can click on the "more reply options" button, and attach photo's directly to the forums.

 

wrong benchmark though. 

[Jumper118]

Intel Core i7 4770k

GTX 780 ti

Core Clock: 1310.7 MHz

Memory Clock: 1749.6MHz 

Avg FPS: 79.5

Score: 3328

Setting: Extreme HD

 

 

 

 

if you overclock your memory then you will get a much higher score.

 

Did I do this right?

 

Intel i5-4690

EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz

74.5

1877

Extreme

 

http://i.imgur.com/RgrbFnK.jpg

 

Intel i5-4690

EVGA GTX 970 SC - 1165Mhz (Max boost 1379.5), 1752.8Mhz

217.7

5485

Basic

 

http://i.imgur.com/A5GwTaR.jpg

that is the wrong benchmark, other than that its fine