GPU Overheating, kernel-power 41, random reboots

[JustRelaxASC]

So in past couple of weeks my PC started randomly restarting by itself while gaming, with no BSOD, screen just goes crazy for a second, I hear a loud noise from speakers and it restarts by itself with no BSOD dump, in Event Viewer it shows Kernel-Power 41, 63 error which basically means nothing to me, I searched on thousands of forums and applied all possible advices and "fixes", sadly with no success.
I've read that kernel 41 error can mean faulty PSU or overheating issue, as I have a relatively new 600W PSU I suspected on overheating
I then took a look at my temperatures and noticed 2 things (Click here to see the image):
1. The GPU temperature reaches over 70°C while under load (gaming)
2. The GPU fans is set to auto in crimson AMD driver software and it doesn't go above ~20% even though GPU is running hot in my opinion, so I tried turning the fans manually to a higher speeds and GPU started coolling down to ~59 meaning that fans can cool it but they don't do that automatically and I don't want to go to crimson every time I start and finish gaming to turn up and down fan speeds as I don't want them to run at 100% while idling.
I also dusted my PC so it's pretty clean and dust is not causing the overheating
Also my GPU never ran above ~50°C while gaming when I bought it.
I'm using MSI Gaming 270X 2GB OC Edition

The Random reboots may not have to do anything with overheating that is just what I suspect as I see no other possible cause at the moment

Any ideas on how to fix the GPU overheating issue?
All help would be much apprechiated

 

[13thmonkey]

70 is not hot. You have a different problem IMO.

Precisely what PSU do you have?

 

[JustRelaxASC]

70 is not hot. You have a different problem IMO.

Precisely what PSU do you have?

Cooler Master B600 600W

 

[13thmonkey]

It's not a great unit.
http://www.hardwaresecrets.com/cooler-master-extreme-power-plus-600-w-power-supply-review/7/
Dies at 600W, out of regulation at 475W.

Unless yours is a bad example it shouldn't cause the problem you are seeing though. But I do urge you to upgrade it.

 

[JustRelaxASC]

It's not a great unit.
http://www.hardwaresecrets.com/cooler-master-extreme-power-plus-600-w-power-supply-review/7/
Dies at 600W, out of regulation at 475W.

Unless yours is a bad example it shouldn't cause the problem you are seeing though. But I do urge you to upgrade it.

I bought it relatively recently and I don't plan on just going all out and buying a new one just because it 'might' be the cause...

 

[13thmonkey]

It's not a great unit.
http://www.hardwaresecrets.com/cooler-master-extreme-power-plus-600-w-power-supply-review/7/
Dies at 600W, out of regulation at 475W.

Unless yours is a bad example it shouldn't cause the problem you are seeing though. But I do urge you to upgrade it.

I bought it relatively recently and I don't plan on just going all out and buying a new one just because it 'might' be the cause...

The 500W seems to be ok.
http://translate.google.com/translate?depth=1&hl=es&rurl=translate.google.com&sl=de&tl=en&u=http://www.planet3dnow.de/vbulletin/showthread.php%3Fs%3D3c1767b253f48bc665e82465083cfaf6%26t%3D413235

PSU's are a remarkably common cause, your temps are no issue, so it is something else. With regards to fan speeds, can you not set a custom curve and leave that on auto?

 

[JustRelaxASC]

With regards to fan speeds, can you not set a custom curve and leave that on auto?

The custom curve in Afterburner wasn't enabled until I enabled it today to see if that helps as fans weren't going over 20% by crimson, so i set them to run at higher speeds on higher temps and temps are now not exceeding ~58 with about 50% fan speed, will see if that helps.

Also I've read that normal temps for this GPU under load is ~65 so that's why I was kinda worried, I also didn't overclocked my gpu at all manually, it's just has a factory OC

 

[13thmonkey]

Undo the factory OC and see what happens.

 

[JustRelaxASC]

How? What should be the normal clock?

 

[13thmonkey]

just reduce by a small amount, 100Mhz to test it.

 

[JustRelaxASC]

just reduce by a small amount, 100Mhz to test it.

I didn't reduce the clock yet, but I set a custom fan curve in MSI Afterbuner like 5 days ago (I set so that they blow at 100% once temp goes over 65) and haven't had crash ever since, so I think it could be that GPU was probably overheating because fans were blowing only 20% being limited by new crimson driver am I right?
But it leaves me wondering why crimson doesn't increase fan speed automatically when it detects that gpu is going over 80°C?

EDIT: Setting custom curve in MSI Afterburner fixed problem for me.

 

[13thmonkey]

Not sure why fans stayed low, but I always use a custom curve.

 

[JustRelaxASC]

Not sure why fans stayed low, but I always use a custom curve.

The problem started when amd driver became so called "Crimson driver", also I haven't been using MSI Afterburner until now, only the basic AMD driver