ASRock H77-ITX + Intel i7-3770 BSOD or crashing

Ragingkileak

Reputable
Feb 19, 2015
2
0
4,510
Hello all,

New poster - hope you can help!

I've recently acquired a cheap, second hand gaming rig which I'm having some trouble with. I'm getting BSOD, hangs or crashes when putting the system under heavy or prolonged load.

The spec (and what I believe to be relevant components are:

Intel I7-3770
ASRock H77-ITX
G.Skill ripjaw 4GBx2

The system initially began to crash when playing games like Rome 2 and CoH, but it wasn't overly frequent. With the arrival of Atilla TW however, it now crashes or hangs in any modest battle.

After a few searches, I found evidence in logs and dumps to suggest that a memory fault was present. I scanned both my C: SSD and G: SATA (where steam lives) and both ok. I then ran Memtedt86 and got 2 clean cycles.

Next I began to run the Intel Diagnostics Tool. This basically places up to 100% load on the CPU. It usually passes through once ok, but on the second concurrent run it will warn me that the CPU may shutdown - which ultimately it does. The system may then take a few moments to restart.

All this is leading me to suspect the CPU is overheating.

My question really is - is that likely?

If so, I don't believe I have overclocked anything (nor do I want to) so what should the settings in the H77 UEFI look like? I understand from research that this mobo doesn't welcome much OC action anyway. The firmware version is 1.9 by the way.

Any help appreciated - a buddy suggested I remove, repaste and reseat the CPU and heat sink.

Thanks
Matt
 
Solution
in general you will not get a hang as a symptom of a overheated main CPU. A "hang" is much more common when a secondary processor like a GPU missed some signal in its electronics. The main CPU has a bunch of checks and will directly call a bugcheck. GPU on the otherhand have to run very fast and have little error checking.

The other symptoms you listed would depend on the actual bugcheck codes and subcodes. It is just hard to tell without looking at the memory dumps.

Generally, the first thing you would want to do in the case of a system hang would be to:
stop overclocking, update the BIOS or reset it to defaults, check the CPU,GPU and power supply fans (blow out dust)

if you have a current nvidia driver, update your ethernet...
Thanks Spentshells, does that sound like a likely fix then?

I have to admit I'm a bit sceptical and although im 'ok' at finding my way around an OS and I've built a PC or two, I have never had to clean and repaste a CPU - I'm worried I'll damage it.

The PC chassis is very small, and the Nvidia card next to the CPU is massive and probably creates a relatively high amount of heat.

Rgds
Matt
 
in general you will not get a hang as a symptom of a overheated main CPU. A "hang" is much more common when a secondary processor like a GPU missed some signal in its electronics. The main CPU has a bunch of checks and will directly call a bugcheck. GPU on the otherhand have to run very fast and have little error checking.

The other symptoms you listed would depend on the actual bugcheck codes and subcodes. It is just hard to tell without looking at the memory dumps.

Generally, the first thing you would want to do in the case of a system hang would be to:
stop overclocking, update the BIOS or reset it to defaults, check the CPU,GPU and power supply fans (blow out dust)

if you have a current nvidia driver, update your ethernet driver (sounds stupid but they install streaming by default and it gets screwed up by old ethernet drivers) Oh, I have even seen nvidia drivers get blocked because of out dated USB drivers. (again, it sounds stupid but I have seen a directx driver chained behind a old stupid USB driver that would not ignore packets that were not for the usb driver, this prevented the directx driver from ever getting a response from the graphics card. 30 seconds later the directx figures the graphics card failed and shut down the system)

(you would have to update the BIOS to get usb fixes, update the CPU chipset to get CPU microcode fixes and usb 2 updates, then update usb 3.0 chipset drivers (external chips) or you would reinstall directx so it was the first in the interrupt chain and the stupid old USB driver would be the last to process the interupt. (USB drivers are slow and should be last anyway)

the point it, memory dumps are useful for figuring out the actual cause of the problems. For this I would run verifier.exe and set some debug flags, then change the memory dump type to kernel or full to allow the windows debugger to use extra debugging commands.

or just update the BIOS, (or reset to defaults) update drivers and remove any overclock software and see if you get another bugcheck. Then post the memory .dmp on a cloud server so it can be looked at.
 
Solution
Yeah that seems like a likely fix, but as you mentioned there is a gpu feeding heat directly to the cpu so you may need additional case fans blowing into the case.

As for cleaning and reapplying paste simply leave the cpu in the socket and be careful not to get the paste all over. Clean the cpu and the heatsink apply a pea sized dot to the cpu then press on the heatsink.