Windows 8.1 locking up intermittently

runswithascript

Hi

I built my computer a few weeks ago and it has been working great although sometimes it locks up completely. This has happened when I am in front of the computer but also when I leave it alone for some time and come back to find it frozen or barely responding before becoming completely frozen. The CPU is an Intel i5-4690K, 8GB of Crucial RAM, I use a gigabyte board, MSI GTX 970, and a Samsung EVO SSD. I am do not overclock any of the components. I use this system almost exclusively for gaming and some BitTorrent downloading using the Deluge client. Reinstalling Windows did not solve the issue. I have tried opening task manager when the issue occurred, to see if something is using up all of the memory, CPU, or reading or writing from the SSD too much, only to receive an error advising it could not be found - task manager worked fine on reboot. Trying to launch RealTemp.exe I received the following error: the application failed to start because its side-by-side configuration is incorrect. Please see the application event log or use the command line too sxstrace.exe tool for more detail'. After that another error appeared with no title on the window: 'open driver handle failure!'. When ever this happens it eventually results in my having to cold reboot the computer.

Have you ever encountered this issue or heard of something similar?

Is there logs somewhere I could check after rebooting to see what is causing the problem?

Outkast_IRE

When trouble shooting this kind of issue i usually check the easy things first, such as RAM.

Do you have any other sticks of ram you can use for a few days to see if the issue clears up or else run a memtest on the ram

runswithascript

Outkast_IRE wrote: »

When trouble shooting this kind of issue i usually check the easy things first, such as RAM.

Do you have any other sticks of ram you can use for a few days to see if the issue clears up or else run a memtest on the ram

Thanks for the reply. No, just the single 8GB Crucial Ballisitx. I will run a memtest.

Any ideas regarding looking at logs, and what to try next if the RAM comes back OK?

runswithascript

Windows 8.1 rebooted into memory diagnostics and found no problem.

Outkast_IRE

Next think i would look at myself would be graphics cards and temperatures.

Do you have any spare graphics card you can use for a few days to see if it stops.

You need to get some temperature monitoring working just to ensure that fans etc are all doing their jobs correctly.

Overheal

Definitely suspecting a hardware issue. One time I had to disassemble an entire rig and rebuild it to fix issues before, can't be too sure something isn't messed up.

Memtest is one thing to do obv. There is also a memory test on theultimatebootcd (google it) you can try. It's also loaded with other diag tools.

Each new build I do gets an overnight memory stress test and I also run prime 95 on it for an hour or two, and 3Dmark's tests (fire strike extreme has been brutal on most builds)

Even though you're not overclockin if something isn't sitting right or is defective than you could be getting low voltage to your CPU or ram.

Capt'n Midnight

http://msdn.microsoft.com/en-us/library/windows/hardware/ff542205%28v=vs.85%29.aspx

you can use truncatememory to ignore the top part of ram as one way

reset the BIOS to system defaults if you've changed RAM

also leave on extended testing overnight

as an aside
in the linux word you can use badram to skip dodgy areas of RAM

in the past people used to suspect PSU's too

also if you live in a rural area / up the mountains lightening damage can happen and it's not good

runswithascript

Firstly, I appreciate all the replies and ideas for resolving this but if someone could come back to me on any Windows software that would allow me to turn on debugging or otherwise check a log after rebooting when the issue occurs it could possibly make this process much faster. Until that time I will look at the other options.

Outkast_IRE wrote: »

Next think i would look at myself would be graphics cards and temperatures.

It has happened when the card is idle in Windows, no game running, and not under load but I will check it anyway. It is a GTX 970, what is a safe temperature, and are MSI's own programs for monitoring the temperature good enough, and should I just keep an eye or stress test it over several hours, if so how many hours?

Outkast_IRE wrote: »

Do you have any spare graphics card you can use for a few days to see if it stops.

Would the i5's on board Intel HD 4600 count?

Outkast_IRE wrote: »

You need to get some temperature monitoring working just to ensure that fans etc are all doing their jobs correctly.

I have CPU_FAN, CPU_OPT, and the three SYS_FANs set through the MSI pp centre >> System information viewer >> Smart Fan Advanced >> to start at 30% power at 30C then scale to 100% power at 60C. Whenever I have looked at this the fans are turning away. The remaining two small 80mm SYS_FANs are set to maximum by the physical fan controller.

Overheal wrote: »

Definitely suspecting a hardware issue. One time I had to disassemble an entire rig and rebuild it to fix issues before, can't be too sure something isn't messed up.

This is something I can do and am planning to at some point anyway, especially since the end of a cable tie dissapeared into it somewhere and I was not able to get it out which has been irritating me. Also, when I start Far Cry 4 and the developers' logos and animations (if you are familiar?) are loading at the beginning the system makes a stranged high pitched noise almost like a beep but not quite. With the case open I was not able to isolate exactly where it was coming from and when I manually set each fan individually (apart from the two graphics) to 100% I was unable to replicate the noise. It does not happen while actually playing Far Cry 4, or on any other game I have tested since, only when the dev's animations are loading.

Overheal wrote: »

Memtest is one thing to do obv. There is also a memory test on theultimatebootcd (google it) you can try. It's also loaded with other diag tools.

Is that memtest any better than the one Windows 8.1 rebooted me into and passed? I can check the other diagnostic tools on UBCD.

Overheal wrote: »

Each new build I do gets an overnight memory stress test and I also run prime 95 on it for an hour or two, and 3Dmark's tests (fire strike extreme has been brutal on most builds)

I had Prime 95 downloaded when I was testing different fan confiruations etc and I wondered if there was a way I could have my system automatically shutdown if the temperature reached say 72C? I am happy to run it for 24 hours then and see what happens, although I have read Prime 95 pushes CPUs too far and other stress tests should be employed for Devil's Canyon processors like mine? It is an i5-4690K. And if fire strike extreme is what I am to use to test the memory over 24 hours is there a way to have the system should down if the memory reaches unsafe temperatures etc?

Overheal wrote: »

Even though you're not overclockin if something isn't sitting right or is defective than you could be getting low voltage to your CPU or ram.

Good to know, the rebuild you mentioned might take care of that.

Capt'n Midnight wrote: »

http://msdn.microsoft.com/en-us/library/windows/hardware/ff542205%28v=vs.85%29.aspx

you can use truncatememory to ignore the top part of ram as one way

Would the RAM passing memory tests, Windows 8.1 (or others?) not suffice?

Capt'n Midnight wrote: »

reset the BIOS to system defaults if you've changed RAM

System only ever had this four week old RAM and I have already reset the BIOS a few times but not siince re-installing, I will do it again for good measure anyway.

Capt'n Midnight wrote: »

also leave on extended testing overnight

Which software would you recommend, and if testing encounters a problem will it point to where it is?

Capt'n Midnight wrote: »

as an aside
in the linux word you can use badram to skip dodgy areas of RAM

Good to know. Ubuntu is my usual operating system, I built this desktop for games which is the only reason I have Windows installed, but when I have overclocking configured and everything else optimal I am going to install Ubuntu and then Mac OS on separate SSDs (I have read with Hackintosh builds this works best).

Capt'n Midnight wrote: »

in the past people used to suspect PSU's too

I paid the premium for the EVGA gold rated 750W but is there a way to test it?

Capt'n Midnight wrote: »

also if you live in a rural area / up the mountains lightening damage can happen and it's not good

This is a possibility, I should probably get a surge protector fitted for the mains to the entire house if that is possible.

Outkast_IRE

The high pitched beep when running a game and developer logos etc come on is actually something i am familiar with, my previous graphics card used to have that noise as well, but it never gave trouble other wise.


Yes swap to onboard graphics for a day or two and see if the same thing happens.

This is the software i use to monitor temperatures across the board http://www.cpuid.com/softwares/hwmonitor/versions-history.html

Just for comparison i have been doing some light gaming tonight and the temps on the board are 27-33 degrees, the processor temps are 35-37 degrees, the graphics card temp is 38 degrees.

ct5amr2ig1nfhp

It wouldn't be something simple like a pagefile problem?

Antar Bolaeisk

Try disabling the on-board NIC for a while and see if it goes away, had a problem on two gigabyte boards that the onboard NIC would cause the computer to completely lock up for a second or two every so often.

runswithascript

Two Y cable fans arrived this morning so I can put all fans through the SYS_FAN headers instead of also using a controller. Now I have them I am going to rebuild completely and see if it makes any difference, perhaps something was loose etc. Update to follow.