GPU Crashes "Fixed" by Underclocking Shaders. - Any fixes to this yet?

The problem
I am sure some of us know that for some reason a few R2 GPU's do not let us overclock, even games at stock cause crashes. The result is that we are forced to force a shutdown by using the power button.

"The Fix"
The "fix" to this has been to underclock the shader clock to 980. Unfortunately this is not a fix in my opinion. I was just wondering if anyone has had any luck. It is something to do with Optimus because no R1 is having this issue.

Further reading can be done Here.

The Question
I was just wondering if anyone has come across something to solve this? It happens across different games (WoW, Black Ops, CSS, etc). So it is not game specific.


Thanks for any thoughts, Sorry for the mess I just threw together,

DeeVu

EDIT: turns out it was a faulty GPU. After four hours on the phone I have a new one on its way.

 

bump.

is my question too wordy?

 

maybe i dont fully understand but sounds like a bad video card to me. if you cant run at least stock clocks (stable) on new or old drivers i cant think of anything else that could be wrong.

 

If you can't run at stock clocks, I don't see why you shouldn't bring your case to Alienware/Dell and ask for a replacement.

 

Well I can confirm that the clock shader is also the bottleneck on my OC. But I don't have to underclock it. I can safely run it at 1110.

I would call dell and ask!

 

I guess ill call up warranty. Would it be able to be replaced onsite or do i have it send it in? I had my screen replaced at home.

 

I have never had to underclock my GPU for games to run...
What games specifically are you having this issue with?

 

Across the board all games that use the Nvidia GPU, it doesnt happen with the integrated card. It has to be a faulty card. Ive already called AW support and after 3 hours have gotten no where on a MoBo replacement. But they will get back to me on monday. Hopefully I can get a crash on video for proof. Just gonna have a one hour gaming sesh and capture one to show the tech thats going to call me back.

 

Yeah that just sounds too ridiculous.. I thought maybe it would be 1 or 2 games but if you getting this on all games defiantly send in for replacement.. strange how it works underclocked though I dont get that part... What temps are you getting just curious?

 

My temps are normal. Thats what baffled the rep. The weird thing is I'm following the idea of the underclock off of that thread I linked as well as one from this forum. It was about 50 page thread on WoW performance issues. It seems as though Im not the only one that needs to underclock shaders by 100mhz to get it to be stable. Its a very weird and seems like very rare occurance. Definitely trying my best to get this replaced.

Edit: just filmed a video of the crash. Basically doesnt let me end the processes through task manager. So I shut down but after the screen turns off the CPU or GPU is still running because you can hear the spinning of the fan. I have to force shutdown. Hopefully this is enough evidence.

QUESTION: should I bring up the fact that underclocking "resolves" the problem? I do not know how AW looks at over/underclocking so I do not know if it'd be wise.

 

Tricky situation.. I guess I would not tell, but then again. A little oc or uc has NOT broke the card.

If u told them, I would say that an "IT freind" of yours told you to do it to check if it still froze, because it was the same problem when he OC'd.

 

This issue is nothing new with nvidia mobile chips. The only difference here is on optimus the process becomes unendable instead of a full system lockup like in the past. Most nvidia mobile cards in the past are not very fond of overclocking and 90% of the time it is the shader clock the sets it off. I am able to get my core near 600mhz and my memory to 900mhz but if i overclock the shaders by any decent amount i run into the same problems (as well as my last 2 9600m gt notebooks).

But if it is not working at stock clocks that classifies as a falty board and you need to get dell to replace it. From what i have learned in the past it is a hardware limitation and there is not software "fix". However non optimus notebooks did get way more milage on the shader clock with much older less optimized drivers but these drivers also put a little less load on the gpu and perform worse.

 

Alright this is alot to think of and you were all very helpful. I will definitely use Phantom's suggestion if it comes to it. Also, thanks Tonka for the information on what's actually happening. I'll report back on monday with any news.

DeeVu

 

After 4 hours of talking on the phone and numerous benchmarks later they have finally cleared me for a new motherboard! So much so that they even put the supervisor on and apologized for the amount of time that had to be taken to complete this task. They are leaving my case open just incase there are still issues following this replacement and it should be replaced by next week

Thanks for all the support/advice everyone
DeeVu

Edit: Update: Got my new MoBo and it seems like all my issues are fixed. I can even OC to 550/1320/890