New rig - random crash

[Remyx]

Hi,

I put together a new rig just a week ago. Most went well, but I experience what seems completely random crashes. I leave the system unattended, comes back to see that it rebooted. It did bluescreen once. The error log shows some services failing (crypto, DNS, IP helper) with event ID 7031, 7032, 7034.

Under Prime and other stress tests, the system doesn't show any sign of instability.

Bios and all drivers have been updated to the latest version.

MB - Gigabyte GA-Z97X-UD5H
CPU - i7 4790K
RAM - Crucial Ballitix Sport XT 16GB
SDD - Samsung 840 EVO
HDD - Seagate 1TB & 1.5TB
no discrete video at this point.

OS - Windows 7 x64

Could this issue be related to disk powering up and down? How do I rule this out?

The motherboard was an open box item. I have few more days to return/exchange it at no cost. How would you further troubleshoot this?

Thank you for any guidance.
Remyx

 

[Reversed]

I would first check the RAM with memtest86 and what PSU are you using?

 

[Remyx]

PSU - Cooler Master RS-750-ACAA-A1

Will run memtest86 right now.


Thank you.

 

[Remyx]

Memtest86 completed sucessfully. No problem found.

Running Prime95 this morning, the rig rebooted after 10 minutes :-(

If the issue comes from the power supply, how would you troubleshoot it?

Thanks.

 

[mfenn]

Can you repeat the quick reboot under prime? A one time occurrence could have been a fluke.

 

[Remyx]

Prime can be reproduced with great regularity. It takes about 7-10 minutes before the reboot.

I installed Intel Extreme Tuning Utility to monitor was was going on.

Interestingly, when Prime stress test is started, the CPU temp goes from idle (30c) to about 68-72c. It stays there for quite a while, looks to be stabilized there under load. But at some point, after 7-10min the temp goes up very rapidly, reaches 90c just before the system reboots.

I can't explain this rather sudden temp rise, which might well be the cause of the system crash/reboot under load.

Note that I have been able to reproduce this only under Prime. Othen stress/benchmark programs have not shown the same pattern.

Remy

 

[mfenn]

90C is the danger zone, so I would say that that's probably your issue. What HSF and case are you using?

 

[Remyx]

It is an old no brand case. Good airflow though.

The HSF is a new Cooler Master Hyper 212 EVO, 4 heat pipes and 120mm fan.

Thanks

 

[mfenn]

It is an old no brand case. Good airflow though.

The HSF is a new Cooler Master Hyper 212 EVO, 4 heat pipes and 120mm fan.

Thanks

Try running with the side panel off and seeing if it still crashes. That'll rule out airflow issues in the case.

I'm suspicious of the case because ~70C is high, but acceptable with a Hyper 212. I'm guessing that the air inside the case might be heating up and it can no longer maintain that temp.

 

[Remyx]

Good guess about the temp. That's also my suspicion that heat is causing the system halt/reboot.

Incidentally, LegitReviews got 71c under load for their test of the 4790K. My results are along the same line. http://www.legitreviews.com/intel-core-i7-4790k-devils-canyon-processor-review_143880/12

All those tests were performed with the side panel on and off. It doesn't seem to affect the outcome.

 

[mfenn]

Good guess about the temp. That's also my suspicion that heat is causing the system halt/reboot.

Incidentally, LegitReviews got 71c under load for their test of the 4790K. My results are along the same line. http://www.legitreviews.com/intel-core-i7-4790k-devils-canyon-processor-review_143880/12

All those tests were performed with the side panel on and off. It doesn't seem to affect the outcome.

Have you confirmed that your Hyper 212 fan is spinning? Does it spin at maximum speed when the CPU gets hot?

 

[Remyx]

Sorry for the late reply, been busy chasing down the issue.

The Hyper 212 fan is spinning, up to max speed under load (2000 rpm).

Here's what I found since yesterday. I might have been mixing 2 distinct issues.

1. An overheating issue has been somewhat widely documented with Z97 MB and 4790K CPUs. This manifests itself violently under Prime95, with CPU reaching >95c within seconds. Other benchmarks and stress tests do not show the same behavior. The issue seems to be related to how the MB supplies current to the CPU under this specific workload. The post on Intel community site covers this extensively.
https://communities.intel.com/thread/54032?start=0&tstart=0

2. My new rig has experienced very occasional crashes/reboot, under no specific workload. To date, I have not been able to reproduce this. These are very occasional and hard to predict.
In my quest to address the problem, I changed the power settings options in Windows 7, preventing the HDD from going to sleep after xx min of inactivity. It might have helped some.
Last night I replaced the MB. Went from a Gigabyte UD5H to a UD3H. So far so good. (the overheating issue and crash under Prime95 is still there

Thanks for all the help.

 

[mfenn]

Interesting, did you try installing XTU as suggested here https://communities.intel.com/docs/DOC-23517 ? Then you'd hopefully be able to see what wattage your CPU was pulling under load.

 

[Remyx]

Yes I did install XTU and watch the car pulling in excess of 120W. This caused immediate temp rise to 100c followed by a reboot.

 

[Ketchup]

I would say that you most-likely need to re-mount your cooler. You should not be getting temps that high with that cooler.

 

[mfenn]

I would say that you most-likely need to re-mount your cooler. You should not be getting temps that high with that cooler.

That's not a bad suggestion. Something simple like forgetting to remove the plastic film or not using any thermal paste could cause a problem. More subtle problems include using too much thermal paste and uneven torque on the mounting screws.