M17x R3 Throttling GPU at 67 Degrees... Someone Help!

As stated, my year old Alienware Mx17r3 Has a 560m GPU, and the GPU has just started throttling at 67 Degrees on the dot everytime, and will continue to throttle, making playing any game or even streaming movies a nightmare.

-I have updated drivers, firmware, and bios.
-reinstalled a fresh windows 7
-taken apart the alienware and cleaned inside and out, new thermal paste.

none of any of this has seemed to help.

Is the card dead in the water, or is there some way to fix this huge bug?

I included a pic where you can see the throttling happening.

 

bump for science

 

kaelib, i've just recently returned to these forums...but when i left, throttling was a very common issue with these cards.

Without doing system changes, the only thing that I can suggest is that you raise the back of your laptop up. You can use bottlecaps, one on either side or use a good working laptop cooler. This should drop your laptop 5-10 degrees which should keep your laptop under the dreaded 67c

Another thing to think about is that sometimes the heatpipes on the heatsink go bad. you can try a new heatsink. This may or maynot help because we don't know if your heatsink is not working properly.

AND

There is also this, written by my friend Widezu:

http://forum.notebookreview.com/ali...throtting-fix-no-need-modify-system-bios.html

 

What drivers are you running? I've been reading a lot of bad things on the most recent Nvidia drivers causing heating issues and killing cards.

 

Yeah re the drivers 314.22 is the last stable WHQL out of nvidia. All the subsequent ones are really bad!

I'm wondering if this is a power throttle rather than overheat. On the 580 the first level of throttle would cut in around that temp but only dropped from 610 to 415. Enough to trash a game though. The overheat throttle was at 76C and the card would drop to 75!! If you use NvidiaInslector you can change the clocks on both power states; P0=full power, P1=power throttle, P8= low power mode, usually while in windows, P12 = help!, I'm in trouble and am going into lowest safe speed (overheat)

Using nvidaInspector to raise P1 to match P0 can get around the problem BUT, P1 also drops the voltage slightly so an OC becomes more difficult. I do not know the clocks or characteristics of the 560 so this may not be applicable but worth checking out

 

Thanks for all your reply's, gonna try all of these suggestions once i get home from work. Ill update the Forum if i can or cant find an answer to my problem in all of this.

 

ok, new data comming in. The P state is staying the same, always at P0 which means it is not thermal throttling. While the video card is running with no cpu load, the GPU load goes from 100% to 0% where it locks up the game, then back up to 100%, then again 0% over and over, ill put up a new pic of this happening. Also, at 100% cpu load, the GPU load will not go above 10-20% and at most times stays at 0%, virtually locking up the game completely. Im at a loss here guys, its not gpu P state throttling, but something new and i cant tell if the cpu is broken, the gpu or something else.




Edit- Also, went ahead and tried using the 314.22 Drivers, no difference, still the same issue, on cpu load, the gpu load goes crazy crashing and loading to 100% in seconds

 

Since you see it in windows this might be worth a run to see the data: View attachment 98122 The DPC number is the one I was told to watch when sound breaks up but it might reveal something. The high hard page faults in the example are because I left the virus checker (MS Essentials) running and it should be disabled to get more accurate data. The DPC number in my example is high (so I have been told) and I know someone that might help if this reveals anything

 

getting ready to run this program, and will post the dpc number and a pic. ty!

on another note, i noticed something new, that may prove that all this is either the sata port on the motherboard or the hard drive itself. While running a test on the hard drive, i was running a read/write test, and during this test i had zero throttles and the gpu load time stayed in the upper 90s without dropping off to 0$ like the other times, as soon as i stopped the read/write test, the GPU load started throttling again from 99 to 0 to 99 to 0 , start back up the read/write on the hard drive, and the gpu load is smooth again, i did this 6-7 times to make sure it was not in my head, and it was not, the hard disk while being forced to work was fixing everything, as soon as it stopped, the throttle/lag came back on. Im almost sure its either the hard drive or motherboard now, if anyone has more experience in this, be greatly appreciated.

 

Unfortunately all that is showing (IMO) is that a drive test suspends whatever process is killing the machine. Depending on the results I will ask one of the big-guns for advice

Did you look in task manager to see if a particular process is popping in as soon as the disk test is completed (sort on CPU usage)? Worth a shot.

 

Have you tried using HWiNFO64?, otherwise id have the same problem, my 580M is a jerk at times, so now i use hwinfo64 and set fans to 4k rpm and BAM cooled, dont get over 70 on high graphic games.