GTX570 overheating, freezing games

[livingsacrifice]

Hmm. Well I did another test and tried a different way I had the PSU setup, still froze but got almost done with 2 Heaven tests and got to a max of 70C temp.

 

[load81]

Overclocking them can blow em up pretty easily and they are lacking in memory.

I wouldn't say they blow up easily. 6 cards on overclock net fried their vrm's by going over 1.1 volts (on stock cooling 200+MHz core overclocks?)or by disabling ocp and running furmark (this killed a few 580's also) and the other 6 in that thread had to be faulty to begin with. That fud thread exagerates every faulty 570 someone gets. Check out newegg reviews faulty cards are everywhere its not just the 570. Not to mention some people in that thread also posted fake 570 deaths. Looking around on the Nvidia forums or anywhere else you don't see 570's blowing up everywhere in the 4-5 months they have been out. Its not so easy to blow them up without editing the bios and disabling safety features. With the abuse I put mine through it would of blown up months ago. And at 1900x1200 and below they have plenty of memory for all but 1 or 2 games.

 

[livingsacrifice]

So here's what I did... as a test I put the clock, mem and shader clocks at the nvidia stock 570gtx speeds (732 core, 1464 shader and 2800 mem) put fan speed at 75% and set unigine on, went and washed the car, came back and it was still running, looks like it did 4-5 complete playthroughs. So I'm willing to bet the "superclocked" settings aren't 100% stable.

 

[OILFIELDTRASH]

I doubt fan speed has anything to do with it. Looks like you have a crappy overclocking card. You could up the voltage a tad and run at the superclocked speeds if you wanted to avoid Rma at the moment.

 

[livingsacrifice]

Works with everything I throw at it with stock speeds, ran crysis 2 for a while, crysis 1 benchmark, unigine heaven and tropical benchmarks. Yep sounds like bad overclocking card to me too.

 

[notty22]

I have a hunch its more the memory clocks. thermal tape and a heat spreader plate over the chips helps cool them. There are many of them and more chance that one gets hot and causes error.
If your doing everything 'right' , this is not your problem.
Return to newegg.
Your best interest is to ask newegg to cross-ship you a new one. Call them and act upset, you may have to allow them to hold a CC number.

 

[OILFIELDTRASH]

You might find a stock clocked 570 plenty for the gaming you do. It's up to you but you did pay for a superclocked version. With evga's outstanding warranty and service you could always RMA at a later date if you want. Evga is pretty optimistic with there overclock. 797mhz isn't a guarantee on stock voltage. Just make sure to register with Evga within 30 days so you get the lifetime warranty.

 

[livingsacrifice]

Agreed, I did pay for superclocked I want that. Running at stock config might as well have just bought a stock card. I think I'm gonna RMA just to be safe, I don't mind waiting I can run most of my games with the 460 just not as smoothly.

 

[livingsacrifice]

In case anyone is still reading this thread, I fixed the issue, was not the card. Memory was set incorrectly and causing it to crash. Guess I should pay attention to the 1.5v sandy bridge requirements.

 

[Idontcare]

In case anyone is still reading this thread, I fixed the issue, was not the card. Memory was set incorrectly and causing it to crash. Guess I should pay attention to the 1.5v sandy bridge requirements.

Interesting. This isn't the first time someone here in the forums has noted that their system ram was the actual root-cause of some GPU instability.

Gotta be frustrating as hell for the AIB's to be incurring the expense associated with supporting their RMA division and to have people returning perfectly good vid cards because the other part of their system, totally outside the control of the AIB, is not setup or functioning correctly.

Consoles don't have this added element of "user induced incompatibility". Might be yet another reason the gaming market is slowly transitioning away from the enthusiast DIY PC segment.

(I'm a perpetrator of this as well, sent an otherwise perfectly functioning ASUS mobo in for RMA only to find out after the fact that it was my CPU that was dying and not the mobo...wasn't ASUS's fault but they bore the bulk of the expense that came from my ignorance and insistence the mobo was the culprit)

 

[livingsacrifice]

Yeah exactly, weird but that was the issue. I really need to swap that ram for some gskill 1600 1.5v so I can run at the ... proper speed.

 

[Makaveli]

Wait what running your ram at 1.65v I assume instead of 1.5v on SB causes GPU instability?

How many reported cases of this is there?

 

[notty22]

I think the problem was , he was running ram rated for 1.6 at the m/b default bios setting of 1.5. Which is the baseline for the memory for SB. But I have heard no problems with going up to 1.65 much like p55 and x58.
Sometimes for the same voltage you can slow your memory speeds down via the multiplier or relax the timings and vice versa.
Increase the voltage and run it a little faster.

 

[livingsacrifice]

What's weird is with those ram settings it was working perfectly fine with the 460gtx. I don't get it, but then again I've seen weirder things over the years.

 

[insurgent]

Interesting, how did you solve this? Was it through accident or have you seen similar problems in another forum?

 

[BFG10K]

What's weird is with those ram settings it was working perfectly fine with the 460gtx. I don't get it, but then again I've seen weirder things over the years.

What's even weirder is that running the GPU at stock nVidia clocks fixed the issue.

So with your system RAM at 1.5v, does the factory overclock on the GPU work properly now?

 

[livingsacrifice]

Yep, been playing on it since yesterday many games for many hours and not one freeze yet. Only 2 things I changed was that ram setting (1.5v at 1333mhz, well in reality I just set the motherboard to AUTO and it did that by itself) and I got a new case that provided better airflow.

And yes gpu is at factory superclocked specs.

 

[nvidiaintelftw]

before you RMA. since many people have said to increase the voltage since the factory overclock does not sound stable, go get MSI Afterburner and enable voltage control and bump the voltage up a bit

 

[livingsacrifice]

I'm not RMAing anymore, I fixed the issue.....

 

[Destiny]

Awesome that you fixed the issue... :biggrin:

Will try the same thing if I start experiencing similar issues...

 

[livingsacrifice]

So... Systems been playing great for the last few days until today. I tried playing around getting my ram up to the proper 1600mhz speed which didn't require an uppage in voltage from the tests I tried. Today though something weird happens. I normally am running EVGA Precision to autoset the fan to 70-75% when playing games so it doesn't get too hot. Well today I was running 3dmark11 after installing the beta drivers (which are awesome btw) and when it gets to one of the cpu or gpu tests like midway through it freezes and closes the window saying theres some sort of error. The really odd part is when it does this it defaults my video card fan back down to 40% but when I open up EVGA precision it is still at 75%. This same issue happened when playing Crysis 2 earlier multiplayer however it didn't close the game just defaulted the fan back to 40%. Why would it be doing this and am I just having an issue with running my memory at 1600mhz at 1.5v?