Solved Occasional BSODs during everyday use

[smiles86]

Hopefully I'm not beating the dead horse, but I just want to make sure. Is this faulty GPU issue the cause for the BSODs?

*Edit: Ran the Furmark test again and paid closer attention. The test doesn't run very smoothly but rather jerks every now and then (pauses and then catches up). The artifacts appear when there is a jerk.

 

[essenbe]

Smiles86, what Paul says is right on. If you have on board graphics, completely remove the card and run with on board graphics for a few days to find out if this is the problem. But, understand, whether this is the cause of the problem or not, your graphics card is malfunctioning. It could be the card or possibly the PSU not delivering the voltage, but the card is not doing right either way.

 

[smiles86]

Ok, I will either remove the GPU or replace it with another and see how things turn out.

For posterity, I ran Furmark one last time for ~5 minutes and observed the same behavior. The test stutters/jerks regularly and artifacts appear when these jerks occur. GPUZ was also run to give an idea of card voltage under load. I admit the first time I ran Furmark I left the computer unattended and returned at the 3 minute mark, so it's entirely possible these artifacts were occurring well before the observed 3 minutes as I had initially stated.

 

[essenbe]

In that screenshot, you GPU temps were at 80C already. With the 600's it will start throttling at 70. But, it shouldn't have been that hot or artifacting. You may want to call tech support for EVGA and I would save that picture for them. I have dealt with them before and they are very good. You should have no problem with them.

EDIT: I may also suggest you download Precision X from EVGA and set a fan profile for your card. Your fan is at 51% while the card is at 80C. The fan speed should be higher. I set a very aggressive fan profile on my card. It will make a little more noise, but will keep temps down. I start at 40 degrees and the fan at 45%. It goes straight up until I'm at 100% fan speed at 65 degrees. The object is stay below 70. That's just what I do, others do it different, but most do have an individual fan profile to control temps.

 

[Layback Bear]

Are the programs being used to do the testing updated to the new Haswell?
I have read many reports some say they have and some say they haven't.

 

[essenbe]

That's true Jack. One of the problems of living on the bleeding edge.

 

[paulpicks21]

In that screenshot, you GPU temps were at 80C already. With the 600's it will start throttling at 70. But, it shouldn't have been that hot or artifacting. You may want to call tech support for EVGA and I would save that picture for them. I have dealt with them before and they are very good. You should have no problem with them.

EDIT: I may also suggest you download Precision X from EVGA and set a fan profile for your card. Your fan is at 51% while the card is at 80C. The fan speed should be higher. I set a very aggressive fan profile on my card. It will make a little more noise, but will keep temps down. I start at 40 degrees and the fan at 45%. It goes straight up until I'm at 100% fan speed at 65 degrees. The object is stay below 70. That's just what I do, others do it different, but most do have an individual fan profile to control temps.

A very good point Steve, 51% Fan speed at 80C is not good, at that temp it really should be 100% fan speed to try and drop the temps.

 

[smiles86]

EVGA has limited hours over the holidays, but I will endeavor to seek an RMA with them ASAP. I will update the status of my system once I have replaced the GPU.

 

[Layback Bear]

Because a Haswell runs hot all by itself under normal use I would suggest installing after market quality cooling any time a Haswell is used. I would also recommend monitor the temps under normal use and see the temp difference between the cores.
Their are many reports of the cores running very different temps because of the method and quality control used to install the heat dissipating cap the heat sink sets on.
It's also my understanding that Intel doesn't consider the problem of different cores running vastly different temps a RMA.
Some have de-caped the cpu and cleaned and installed new thermal compound and solved the problem. Not for the faint at heart to attempt. If just one of the cores gets to hot it will throttle down.

I can't post all the sites I did my research because it was a while back when I was doing my research for my new system and yes I did a lot of research.

In my opinion unless one is lucky or Intel changes its process of installing the heat dissipating cap on the cpu over clocking is questionable at best because of the core temps.
Loading a system using programs (i.e. Prime 95, AID, ect) is not recommended by Intel or Asus because it might damage the cpu and possibly the motherboard.
Because their are so many (IF's and Maybes) is the reason I didn't choose the Haswell.
If it was my computer I would run everything OOTB fashion with a quality after market cpu cooling.
**It will make diagnosing BSOD more difficult.

PS: Make sure all plastic is removed from the video card. Some are very hard to see and will block air flow if left on the video card.

Note: Some Intel Haswell did very well in testing from the factory but most Retail Haswell didn't do so well. I'm guessing the Intel did a little cherry picking when the sent out testing cpu's.

 

[paulpicks21]

Because a Haswell runs hot all by itself under normal use I would suggest installing after market quality cooling any time a Haswell is used. I would also recommend monitor the temps under normal use and see the temp difference between the cores.
Their are many reports of the cores running very different temps because of the method and quality control used to install the heat dissipating cap the heat sink sets on.
It's also my understanding that Intel doesn't consider the problem of different cores running vastly different temps a RMA.
Some have de-caped the cpu and cleaned and installed new thermal compound and solved the problem. Not for the faint at heart to attempt. If just one of the cores gets to hot it will throttle down.

I can't post all the sites I did my research because it was a while back when I was doing my research for my new system and yes I did a lot of research.

In my opinion unless one is lucky or Intel changes its process of installing the heat dissipating cap on the cpu over clocking is questionable at best because of the core temps.
Loading a system using programs (i.e. Prime 95, AID, ect) is not recommended by Intel or Asus because it might damage the cpu and possibly the motherboard.
Because their are so many (IF's and Maybes) is the reason I didn't choose the Haswell.
If it was my computer I would run everything OOTB fashion with a quality after market cpu cooling.
**It will make diagnosing BSOD more difficult.

PS: Make sure all plastic is removed from the video card. Some are very hard to see and will block air flow if left on the video card.

Note: Some Intel Haswell did very well in testing from the factory but most Retail Haswell didn't do so well. I'm guessing the Intel did a little cherry picking when the sent out testing cpu's.

But apart from that it's a good CPU right? :roflmao:

I have to say, I built a rig with the same CPU a couple of months back for a family member, he is a heavy gamer and has not had one issue with it even using the stock cooler.

 

[smiles86]

If it was my computer I would run everything OOTB fashion with a quality after market cpu cooling.

I agree, I should definitely buy an aftermarket CPU cooler. Those temperatures I experienced during Prime95 really concern me. As far as temperature variation between cores go, I've not personally noticed anything greater than ~5 C max difference.

PS: Make sure all plastic is removed from the video card. Some are very hard to see and will block air flow if left on the video card.

I'm not sure what plastic you refer to that would be on the video card, but I took a look at it and didn't spot anything besides the plastic shroud that's part of the card. The blower does not appear to be obstructed. I did have the SLI cap left on since I only have one card.

 

[Layback Bear]

smiles86 sometimes the metal plate on the opposite side of fan mounting has plastic and some times the mounting end has plastic on a vent. Just a caution.

To my understanding 5 deg difference is no problem.
On my i960 with a lot of lapping I got the difference down to 3 deg. Haven't tried it on my new chip as of now.

 

[Layback Bear]

paulpicks21 I'm not saying the Haswell isn't a good cpu. In my opinion it's not one to be selected to do over clocking.
Because they run hot normally like my i960. Better cooling in my opinion should be used even running stock clock. Can't hurt anything and might help.

 

[smiles86]

I had the GPU replaced and also installed an aftermarket cooler on the CPU. Been running for about a month now without any incidents, so I think we can call this case solved. Many thanks for the help.

 

[essenbe]

Well done. I'm glad you have it running like you want it. Good luck and let us know if you need anything further.

 

[Layback Bear]

Very good smiles86.
Could you let us know what cooler you chose for your cpu and what temps you are getting?