GTX670 causing random shut downs?

[webbo254]

Hi guys

I've been having this problem for around 6 months now and have still not found a solution.

Specs:

i5 2500k @ 4.2GHz

Palit GTX 670

MSI P67A-GD53(B3)

8GB Kingston Ram

OCZ Agility 3 60GB SSD

2 x Samsung Spinpoint HDD's

The problem is that at random times (usually once every 10 days or so) my pc will just randomly cut out and after a few seconds turn itself back on again just like somebody flicking the power switch on the back of the PSU and turning it back on again. The fact that it happens so infrequently, can go 3 or 4 weeks sometimes without it happening makes it very annoying to make changes and test. I am 100% it isnt a power or heat issue as I have experience this issues with two different power supplies, and I am always monitoring my temperatures and the CPU never goes above 60-65 and the GPU never goes above 70 degrees.

Things I have tried:

-My first thought was the PSU so replaced my CoolerMaster Silent Pro M 700w with an XFX Core 650W.

-Updated Motherboard BIOS

-Taken motherboard out, checked all connections and reseated everything.

-Removed GPU overclock

-Run MemTest and passed with no errors

One thing I have noticed is that it has never happened while gaming, it is always while just browsing the net or doing some work. This is another reason I dont believe it is a power or heart issue.

After it happens in the event viewer I just have a kernel power 41 error.

Hopefully somebody can suggest something else to try?

Thanks

[Opsilon]

Ok m8, most ppl ignore the fact that cpu stability testing has to be done at least 6 hours but the longer you do it the better so 24-48 hours should be best, because you needed to change voltage for 4.2 on Sandy Bridge, and most ppl just run it for 3-6 hours and say it's 24/7 stable, so if you did not test it for 24-48 hours, try puting your cpu to stock speeds and voltages and try it out, if you did that ignore what i said and do the following, oh 1 more thing do not lie to yourself that you did 24-48 h testing if you did not, don't be lazy...

1. try reinstalling windows

2. did you run some SSD tests, is suspect you run your windows of it

3. do you play some games of the ssd? ( i know it's a small one, but have to ask )

4. do you have a UPS?

5. do you live in an old building with old power installations

6. did you try it with another GPU? because gpu-s would crash more often when using more power and when things get "rough"

7. I had a similar problem with my old pc, i had windows and basic stuff on 1st HDD, and all programs and games on the 2nd, and it turned out 1 of them was failing

8. Try installing windows on your samsung drive, test it out for a few days

[webbo254]

Ok m8' date=' most ppl ignore the fact that cpu stability testing has to be done at least 6 hours but the longer you do it the better so 24-48 hours should be best, because you needed to change voltage for 4.2 on Sandy Bridge, and most ppl just run it for 3-6 hours and say it's 24/7 stable, so if you did not test it for 24-48 hours, try puting your cpu to stock speeds and voltages and try it out, if you did that ignore what i said and do the following, oh 1 more thing do not lie to yourself that you did 24-48 h testing if you did not, don't be lazy... 1. try reinstalling windows 2. did you run some SSD tests, is suspect you run your windows of it 3. do you play some games of the ssd? ( i know it's a small one, but have to ask ) 4. do you have a UPS? 5. do you live in an old building with old power installations 6. did you try it with another GPU? because gpu-s would crash more often when using more power and when things get "rough" 7. I had a similar problem with my old pc, i had windows and basic stuff on 1st HDD, and all programs and games on the 2nd, and it turned out 1 of them was failing 8. Try installing windows on your samsung drive, test it out for a few days [/quote']

Thanks for your reply.

I have the CPU overclocked on turbo so i left voltages on auto so it just increases speed when needed. I have never run a long stability test just guilty at thinking it is because it has always been fine while gaming.

1.) I havn't tried reinstalling windows yet,.

2.) What are the best methods for testing SSD's?

3.) No as it is only small only Windows and some programs are installed on the SSD, all games are run off the HDD's

4.) No just a surge protector, i have tested it plugged in to that and straight in to the wall and has happened on both.

5.) I have actually moved house recentely and it has happened multiple times in both locations, my new house being new.

6.) I dont have another GPU test, however this has only ever happened while idle, never done it while gaming or doing anything intensive.

7.) All of the windows files are on the SSD, I made sure to only have the SSD connected when I installed windows.

[mgsstar]

Ok m8' date=' most ppl ignore the fact that cpu stability testing has to be done at least 6 hours but the longer you do it the better so 24-48 hours should be best, because you needed to change voltage for 4.2 on Sandy Bridge, and most ppl just run it for 3-6 hours and say it's 24/7 stable, so if you did not test it for 24-48 hours, try puting your cpu to stock speeds and voltages and try it out, if you did that ignore what i said and do the following, oh 1 more thing do not lie to yourself that you did 24-48 h testing if you did not, don't be lazy... 1. try reinstalling windows 2. did you run some SSD tests, is suspect you run your windows of it 3. do you play some games of the ssd? ( i know it's a small one, but have to ask ) 4. do you have a UPS? 5. do you live in an old building with old power installations 6. did you try it with another GPU? because gpu-s would crash more often when using more power and when things get "rough" 7. I had a similar problem with my old pc, i had windows and basic stuff on 1st HDD, and all programs and games on the 2nd, and it turned out 1 of them was failing 8. Try installing windows on your samsung drive, test it out for a few days [/quote']

Thanks for your reply.

I have the CPU overclocked on turbo so i left voltages on auto so it just increases speed when needed. I have never run a long stability test just guilty at thinking it is because it has always been fine while gaming.

1.) I havn't tried reinstalling windows yet,.

2.) What are the best methods for testing SSD's?

3.) No as it is only small only Windows and some programs are installed on the SSD, all games are run off the HDD's

4.) No just a surge protector, i have tested it plugged in to that and straight in to the wall and has happened on both.

5.) I have actually moved house recentely and it has happened multiple times in both locations, my new house being new.

6.) I dont have another GPU test, however this has only ever happened while idle, never done it while gaming or doing anything intensive.

7.) All of the windows files are on the SSD, I made sure to only have the SSD connected when I installed windows.

Did you try benchmarking with Heaven/3DMark and see if there are any artefacts in the video?

[webbo254]

Yes I have tried a number of benchmarks with no issues and they all pass fine.

[ProGearUk]

Are you sure that it is not a problem with the gpu becoming unseated. I have a file server with a AMD Radeon HD 5450 in it and that came unseated from the motherboard and caused the whole computer to turn off. Maybe it is that. Ensure all your connections are in properly.

[webbo254]

Are you sure that it is not a problem with the gpu becoming unseated. I have a file server with a AMD Radeon HD 5450 in it and that came unseated from the motherboard and caused the whole computer to turn off. Maybe it is that. Ensure all your connections are in properly.

I have checked all my components multiple times and reseated the graphics card at least 4 times.

[AJRicho]

default / reset you bios. check your ram settings and make sure correct to the ram. But this could be related to the network card. Is the network on board or pci express/pci ??