Weird issue overclocking my GTX570.

[MercuryDoun]

So Heres something weird... I can run my GTX570 @ 970/1940/2420 through 3DMark11 with 1125mv, runs great! Got a good score too, lol... But the EXACT same speeds(970/1940/2420) crashes my computer out right at 1150mv during the 3rd or 4th test...

 

The exact same speeds again at 1213mv crashes almost instantly, sometimes the first test doesnt even fully load...

 

The only think i can think of that might be the problem is that the card isnt getting enough juice, or more to the point, enough amps. Here is the PSU i currently have: http://www.newegg.com/Product/Product.aspx?Item=N82E16817371024 or http://www.antec.com/Believe_it/product.php?id=NzIw

 

It has 4 12v rails @ 22A/22A/25A/25A. During testing i have my CPU at 4.0Ghz, but its only at 1.25vcore.

 

I dont think its my my GPU core itself, typically while overclocking a video card a total system crash means not enough voltage correct? Its tearing and artifacting that means the core is at its limits regardless of voltage yes?

 

Thanks in advance for any responses!

 

I wouldnt be opposed to getting another PSU 850w+ with a single strong 12v+ rail if it could/might resolve the problem. But im reluctant to do such as thing unless i have a pretty darn good reason to believe the PSU is truly at fault. As well, it would be EXTREMELY difficult to sell this PSU due to its CPX form factor. It only fits in a handful of Antec cases and thats it. I only paid like $80 for the thing in the first place due to a UBER combo deal at the egg.

[El_Capitan]

Try rebooting? Sometimes if the GPU fails and you just reload the same settings in MSI Afterburner, it crashes more often.

 

If it's tearing and artifacting, that doesn't necessarily mean that the core is at it's limit. Sometimes you can give it more juice and it will stop. However, it can only take so much juice before something blows.

 

The only way to tell if it's your PSU is to measure the current directly from the rails. 850W's should be fine, though.

[MercuryDoun]

Try rebooting? Sometimes if the GPU fails and you just reload the same settings in MSI Afterburner, it crashes more often.

 

If it's tearing and artifacting, that doesn't necessarily mean that the core is at it's limit. Sometimes you can give it more juice and it will stop. However, it can only take so much juice before something blows.

 

The only way to tell if it's your PSU is to measure the current directly from the rails. 850W's should be fine, though.

 

Ive rebooted several times outside of the crashes, so i dont believe that to be the problem. And i havent actually seen in tearing or artifacting yet. Just crash, reboot lol.

 

I know 850w's is more than good, but what if both pci-e 6pin power connectors are pulling from the same rail... That could potentially be under powered then yes, with only 22a or 25a? Im 95% certain that i made sure one of the 6-pin connectors was from the non-modular portion of the PSU and the other 6-pin was from the modular part. So in theory they should be on separate rails, i guess it could all depend on the innards of the PSU though.

[MercuryDoun]

Bump! Looking for more opinions!

[El_Capitan]

Try putting it in another PCIE2.0 slot or up the QPI voltage or maybe the ICH PCIE Voltage a notch?

[MercuryDoun]

Try putting it in another PCIE2.0 slot or up the QPI voltage or maybe the ICH PCIE Voltage a notch?

 

This i will try when i get home.

[paulktreg]

I doubt very much you've hit the current limit on the rails. The 25A rails, 12V3 and 12V4, will be used for the PCI-E connectors and even if you've got 2 x 6 Pin PCI-E on the same rail don't forget they're only rated at 75W or 7A each.

[MercuryDoun]

I doubt very much you've hit the current limit on the rails. The 25A rails, 12V3 and 12V4, will be used for the PCI-E connectors and even if you've got 2 x 6 Pin PCI-E on the same rail don't forget they're only rated at 75W or 7A each.

 

That makes sense... This is the kind of advice i was looking for! Ill really have to play with some BIO's settings to get things off on the right foot then. I didnt think about a single 6-pin connector only using 7A each.

 

Thanks Capi and Paul!

[wevsspot]

MD - what happens if you set MSI Afterburner to use a constant voltage? Same results?

[Nyt]

Buy an AX1200 that has a single rail then problem solved

[MercuryDoun]

MD - what happens if you set MSI Afterburner to use a constant voltage? Same results?

 

 

I didnt know i could do this? During testing the voltage, according to HWmonitor, does jump to the proper amount though. Im thinking Capi might be onto something with my ICH or QPI voltage.

 

Ill definitely set AB it use constant voltage and see if that helps! Thanks for the tip.

 

Buy an AX1200 that has a single rail then problem solved

 

Lol, Yes. Because i have $260 just laying around.

[wevsspot]

No harm No foul bumping the QPI and ICH voltages up a bit. That might do the trick.

 

I was just wondering if maybe the voltage transitions were creating some problems at certain set point voltages. May or may not be creating the issue. Only testing will tell