New 3DMark released.

[Shinzai]

Well, no not really. I could push it through Heaven at 1400+ core and 7800+ memory.
Perhaps mild wasn't the best description but I just pulled down the sliders a fair bit from my old stable overclock i used with the older drivers and Windows 7. With the free version of 3DMark I just don't have the patience to find max overclock since it only fails in the last test and it takes way too long to reach.

Interesting, I didn't know the windforce could stay stable at such high clocks.

[n0cturnal]

Interesting, I didn't know the windforce could stay stable at such high clocks.

I don't think it is a general attribute for the card, I just got lucky with the silicon.
Straight out of the box the Kepler boost puts my card at 1215Mhz.

[Faithh]

Oh yeah I know it is coil whine, I think I formulated that sentence a bit wrong. We were talking about my bad results for the first test, that's what I don't understand.

Compare my overclocked score versus yours, for the 2nd and 3rd test they aren't that far apart but the first test is a massive difference.

Well I just got an idea for you, you might like it :P Can't you hotkey an overclock profile to your keyboard with msi AB? So just run at stock settings TILL you see the loading screen of firestrike and push that hotkey?

Actually Firestrike, especially graphics score is the only interesting thing for us.

Btw have now 8612 on graphics score in FS. http://www.3dmark.com/3dm/24309

[n0cturnal]

Well I just got an idea for you, you might like it :P Can't you hotkey an overclock profile to your keyboard with msi AB? So just run at stock settings TILL you see the loading screen of firestrike and push that hotkey?

Actually Firestrike, especially graphics score is the only interesting thing for us.

Btw have now 8612 on graphics score in FS. http://www.3dmark.com/3dm/24309

That doesn't really help, I still have to sit through it all. I'll do a proper max someday I just don't have the patience to do it now :P

GTX 680 is totally cheating

[Faithh]

That doesn't really help, I still have to sit through it all. I'll do a proper max someday I just don't have the patience to do it now :P

GTX 680 is totally cheating

Idd :P But your card is still amazing..

Something else I hate about this Asus card is I can't get the vcore to 1.175V during full boost clock, it's just sitting at 1.14V orsomething only while I've set 1.175V >.< Tried to bios mod a month or two to 1.215V but was again only sitting at 1.185V

[n0cturnal]

Idd :P But your card is still amazing..

Something else I hate about this Asus card is I can't get the vcore to 1.175V during full boost clock, it's just sitting at 1.14V orsomething only while I've set 1.175V >.< Tried to bios mod a month or two to 1.215V but was again only sitting at 1.185V

I was under the impression that only MSI PE could change the voltage.
Googled a bit and found an unlocked BIOS for my card, this is tempting.

[Faithh]

I was under the impression that only MSI PE could change the voltage.
Googled a bit and found an unlocked BIOS for my card, this is tempting.

Ye If I can give you some advise with flashing your bios, just make sure you make a backup of your default bios through nvflash NOT with gpu-z. The backup bios from gpu-z was unable to be flashed through nvflash so basically nvflash was telling me the .rom file had an internal IO error and didn't flash. I just tried to flash the backup bios with the bios on the chip already to make sure it's working before going to mod the bios.

Higher than 1.215V is not possible because there's a resistor on newer cards. If you can't find a bios mod for your card, you can do it with kgb tool -> http://www.xtremesystems.org/forums/...ditor-Unlocker

There were also different revisions for my gtx 680 top and the premodded caused loads of artifacts so that's why I had to KGB.

[n0cturnal]

Ye If I can give you some advise with flashing your bios, just make sure you make a backup of your default bios through nvflash NOT with gpu-z. The backup bios from gpu-z was unable to be flashed through nvflash so basically nvflash was telling me the .rom file had an internal IO error and didn't flash. I just tried to flash the backup bios with the bios on the chip already to make sure it's working before going to mod the bios.

Higher than 1.215V is not possible because there's a resistor on newer cards. If you can't find a bios mod for your card, you can do it with kgb tool -> http://www.xtremesystems.org/forums/...ditor-Unlocker

There were also different revisions for my gtx 680 top and the premodded caused loads of artifacts so that's why I had to KGB.

Yeah this seems to be a bit risky, need to read up on it before I even touch this.

[Yohassakura]

7970 reporting in...

http://www.3dmark.com/3dm/25446

Says it doesn't recognize the driver, but I'm running 13.1 (official).

[Shruikah]

7970 reporting in...

http://www.3dmark.com/3dm/25446

Says it doesn't recognize the driver, but I'm running 13.1 (official).

Crossfire 7970's here. I had a fair bit of artifacts happening during the cloudgate one. Using the same driver version as you. Stock speeds aswell.

http://www.3dmark.com/3dm/25694?

[Yohassakura]

Crossfire 7970's here. I had a fair bit of artifacts happening during the cloudgate one. Using the same driver version as you. Stock speeds aswell.

http://www.3dmark.com/3dm/25694?

Hmm, something definitely went wrong with the 2nd test (cloud gate) if I beat you there... And I'd expect a bigger difference between 1 OC'd card vs 2 stock cards.

(Mine is 1200 / 1600 in comparison with your 1000 / 1450).

Sounds like it could have been not enough power, or over-heating - though you did well in the others.

[Shruikah]

Hmm, something definitely went wrong with the 2nd test if I beat you there... And I'd expect a bigger difference between 1 OC'd card vs 2 stock cards.

(Mine is 1200 / 1600 in comparison with your 1000 / 1450).

You also have a 3770k at 4.5 - look at the graphics score

[Yohassakura]

You also have a 3770k at 4.5 - look at the graphics score

Oh... Derp. Didn't see much CPU usage. (Core 4 + 6 were doing nothing... or 3 + 5 if you start from core#0)

You should have had a much higher graphics score on #1 then, surely...

[Deleted]

Hmmm, could be that 3dmark still doesnt play nice with 2+ cards.
Downloading it now, will post results later

Results, with a stock 7970

http://www.3dmark.com/3dm/38499

Did look ugly tho, probably because its at 720/1080p and my screen is 1440p. And there seems to be something weird with my 2500k, that should be at 4.5 GHz, instead of the 1.7

[Blaze611]

GTX 670 SLi coming in. 1054 MHz (Non-Boost) 1228 MHz(Boost Clock) 3310 MHz VRAM.

http://www.3dmark.com/3dm/48148?

[Majesticii]

@Nocturnal the card screaming is pretty normal for any 600 or 7000 series. It's not PSU related or anything related like that but it's just ONLY happening when you really have absurd high FPS, which is coil whine. Not that I know perfectly why it's coil whining but there's a lot of frequency switching shit happening in VRM on the pcb of gpu and causing the coils vibrating.

When you launch a very old game like Halo Combat Evolved, this card is pretty much boosting to 2000+ fps but once you enable vsync @60fps the screaming noise disappears complety.

Score with a gtx 680 - 1300MHz @ 1.35V - 7560MHz on VRAM @ 310.90 drivers -> http://www.3dmark.com/3dm/23107

Read other posts first, before commenting please. High fps causes maximum power draw, which makes your onboard powersupply (VRM's) work at their max. I already looked at the boards, and the gigabyte has a reasonably modest power supply of 5vrm's. Compared to f.e. a ASUS DCII which has 8VRM's. Probably the high fps caps off earlier on his card due to the ONBOARD (NOT MAIN PSU) power supply.

So this is not something only common to GTX600 boards. Maybe because the refference boards of the GTX600 series are very small and have even more limited power supply, this may occur more often in comparisson to other PCB's.

Benchmarks with a non-retarded fps range the 5 phase seems to be sufficient, even at higher clocks. Which makes me question the purpose of the first test. Is it designed to test PCB quality, or just really wierd choice.

[Deleted]

Benchmarks with a non-retarded fps range the 5 phase seems to be sufficient, even at higher clocks. Which makes me question the purpose of the first test. Is it designed to test PCB quality, or just really wierd choice.

If you mean the 3 tests with 3dmark. The first one is aimed at tablets and phones, the second one at laptops and your everyday computer, and the last one for gaming machines. So actually, only the last test is of importance for us :P

[Majesticii]

If you mean the 3 tests with 3dmark. The first one is aimed at tablets and phones, the second one at laptops and your everyday computer, and the last one for gaming machines. So actually, only the last test is of importance for us :P

aha, well, by accident the test is a nice quality check for graphics card PCB boards aswell now :P

[Deleted]

Well, i got coilwhine with my 7970 as well -.- Bloody annoying. Happens in some game menu's as well.

Come to think of it. It is also the reason you only see a real improvement in graphics score at the last test when using 2 cards. The other 2 tests dont need it.

[Majesticii]

Well, i got coilwhine with my 7970 as well -.- Bloody annoying. Happens in some game menu's as well.

I only get a whine in that test, and it's not very loud. But since my case is right next to me, and open mesh, i can easily hear it.
ASUS boards are pretty high quality i must say, atleast this one is.