Sign in with
Sign up | Sign in
Your question
Solved

PC crashes when under high load?

Last response: in CPUs
Share
December 26, 2013 12:23:47 AM

Hello all. I recently build a custom gaming pc and it works great for everything accept that, when it is under extreme cpu/gpu loads it blue screens and says hardware failure bla bla bla, but it restarts and runs just fine after. I was rendering a scene in blender and the cpu went to 100% usage and was doing a good job and then it froze for about 1/2 i sec and then blue screen. I was also playing crysis 3 and i went to change graphic settings* click, flick, ultra ultra,* done! apply! *freeze* crash, blue screen annoying sound from sound card.

My specks are as follows.
Asus gtx 770
intel i7 4770k (non oc)
asus maximus VI hero mobo
6 fans, and a fan controller using a molex.
corsair HX1050 (single 12v rail)
1 ssd
1 hdd
1 disk drive
2 chips of ddr3 memory
i forget what kind of cpu cooler i use but its good and temps are rarely goes over 35 degrees C.


so, what is my problem, its kinda annoying me knowing something is not right, i will be happy to post any more info if need be.

Cheers - A confused first time pc builder.

More about : crashes high load

a b à CPUs
December 26, 2013 12:39:19 AM

The setup looks ok so need to eliminate components, What are your temps at full load on both GPU and CPU. Make sure all drivers are up to date. First run Prime95, if the PC freezes then its the CPU causing the problem (run for 30mins-1 hour monitoring CPU temp). If ok run something like MSI combuster to test the GPU (again monitor the temps). seems long winded but its just eliminating components.
a c 147 à CPUs
December 26, 2013 12:48:52 AM

mmm... not enough info to take a great guess at the problem, but i can take several.

blue screen errors like you're getting are often hardware driver related.
1) I would update the bios on your motherboard, and go to the ASUS website. make sure you get all the latest drivers for your motherboard, and install them.
2) Then i would uninstall the gpu drivers for your video card... uninstall everything you suspect is nvidia related, restart your system, then download the latest WHQL drivers and do a fresh install of your video drivers. I would suggest you do a custom install and NOT install nvidia experience as i've heard nothing but horror stories attached to that software, so we'll make sure it's not the issue and just not install it.

if this doesn't get your problem you'll have to do a bit more detective work... but my next guess would be you've got faulty ram... so i would download memtest86, load it onto a thumb drive and boot to the thumb drive. you'll need to pull out one stick of ram before you run it, then run it for a dozen passes or so on your one stick of ram. turn off the system and try it with the other. Then try them together. If it comes up clean then your ram isn't the issue.

At this point i would suspect you've got temp issues, first we'll eliminate the motherboard and gpu overheating... pop the side of your case off and stick a room fan in the opening. if you still are getting the crashes, we'll move onto the cpu cooler. pull it off, clean the cpu and bottom of the cooler with rubbing alcohol and qtips, reapply your thermal paste. follow the instructions carefully (there are a lot of guides for how to do this, i suggest you use the "pea" method).

if at this point you're till having an issue i would suspect it's the video card or power supply. and swap each out with older hardware that you know works, one at a time... and see if the issue persists.

ideally you'll have found the issue by this point. let me know how it goes.
Related resources
December 26, 2013 1:03:01 AM

Well, I would venture a guess to say its not my drivers. I just built this PC and used all the current drivers. I plan to run prime 95 and collect some data if it doesn't crash. And I would like to stress that this has only happened 3 times over the past 6-7 months after I built it. Everything else runs fine.
a c 147 à CPUs
December 26, 2013 1:13:50 AM

Julian Cecchetto said:
Well, I would venture a guess to say its not my drivers


well considering 95% of windows blue screen errors is caused by a driver conflict or driver failure, i wouldn't be so quick to dismiss driver issues.

if you want to solve this problem follow the suggestions i posted. if you want to deny all the possibilities before testing them, then why ask for a solution in the first place? start with the temp suggestions then... those are also plausible from what you describe. of course only 3 blue screens in 6 months isn't exactly a repeatable issue. unless you can blue screen in prime95 consistently, figuring out this problem will be hard.
December 26, 2013 8:53:06 AM

if you get a bluescreen, you can take a stab at the problem using whocrashed.exe on your memory dump file. Also if you post the memory dump file on
the a cloud service like skydrive. pepople can take a quick look at it with a debugger and see what the OS thinks the problem is.

I have seen problems like this caused by issues in firmware in SSD. I am not saying that is your issue but see if you have a firmware update for your ssd. Also, you can often work around the issue if you go into BIOS and set your SATA port for your SSD to be hotswap enabled.

Sorry to give a "solution" with out looking into the actual cause.
December 26, 2013 9:56:37 AM

So, i have just started Prime95 set to combined, 8 tests and it crashed to blue screen less then 5 minutes into the test, about 2-3. should i do a cpu only tests, rather than combined?
December 26, 2013 10:01:11 AM

How do i find the data logs of windows blue screens and prime 95
December 26, 2013 10:02:06 AM

Julian Cecchetto said:
So, i have just started Prime95 set to combined, 8 tests and it crashed to blue screen less then 5 minutes into the test, about 2-3. should i do a cpu only tests, rather than combined?


check your memory timings, or underclock your memory and rerun the test and see if it runs longer.

December 26, 2013 10:17:17 AM

Well, i have taken 1 of my ram sticks out and run the moderate ram test (second one) (above combined) torture test, and it has run the longest of any test without crashing, if i leave this run for say, 24 hours can we be fairly sure this ram stick is the problem, or am i off in wonderland here.

UPDATE: i would like to not touch OC settings on this pc because i do not know what i am doing with over clocking and i left everything at stock clocks.
December 26, 2013 10:26:24 AM

Julian Cecchetto said:
Well, i have taken 1 of my ram sticks out and run the moderate ram test (second one) (above combined) torture test, and it has run the longest of any test without crashing, if i leave this run for say, 24 hours can we be fairly sure this ram stick is the problem, or am i off in wonderland here.

UPDATE: i would like to not touch OC settings on this pc because i do not know what i am doing with over clocking and i left everything at stock clocks.


sounds ok, but do note that many BIOS updates are done because the BIOS has the incorrect timings set as defaults in their firmware. Check the BIOS update list for your motherboard, or you may be working on a known/fixed problem. Also some RAM slots will require different timing settings for secondary banks even with the same RAM type in both slots. Some motherboards will often specify this in the fine print with * in their motherboard manual. I would hope the BIOS would adjust the timings but often they don't or you have to update the BIOS to get "memory compatibility fixes"

or heck you just might have a bad stick of ram
December 26, 2013 10:38:44 AM

johnbl said:
Julian Cecchetto said:
Well, i have taken 1 of my ram sticks out and run the moderate ram test (second one) (above combined) torture test, and it has run the longest of any test without crashing, if i leave this run for say, 24 hours can we be fairly sure this ram stick is the problem, or am i off in wonderland here.

UPDATE: i would like to not touch OC settings on this pc because i do not know what i am doing with over clocking and i left everything at stock clocks.


sounds ok, but do note that many BIOS updates are done because the BIOS has the incorrect timings set as defaults in their firmware. Check the BIOS update list for your motherboard, or you may be working on a known/fixed problem. Also some RAM slots will require different timing settings for secondary banks even with the same RAM type in both slots. Some motherboards will often specify this in the fine print with * in their motherboard manual. I would hope the BIOS would adjust the timings but often they don't or you have to update the BIOS to get "memory compatibility fixes"

or heck you just might have a bad stick of ram


Ok, so, look in bios update notes, if there is one, and if that is inconclusive compare the recommended clock of the memory to the clock set in the bios? i have asuses v bios.

Best solution

December 26, 2013 10:56:35 AM
Share

Julian Cecchetto said:
johnbl said:
Julian Cecchetto said:
Well, i have taken 1 of my ram sticks out and run the moderate ram test (second one) (above combined) torture test, and it has run the longest of any test without crashing, if i leave this run for say, 24 hours can we be fairly sure this ram stick is the problem, or am i off in wonderland here.

UPDATE: i would like to not touch OC settings on this pc because i do not know what i am doing with over clocking and i left everything at stock clocks.


sounds ok, but do note that many BIOS updates are done because the BIOS has the incorrect timings set as defaults in their firmware. Check the BIOS update list for your motherboard, or you may be working on a known/fixed problem. Also some RAM slots will require different timing settings for secondary banks even with the same RAM type in both slots. Some motherboards will often specify this in the fine print with * in their motherboard manual. I would hope the BIOS would adjust the timings but often they don't or you have to update the BIOS to get "memory compatibility fixes"

or heck you just might have a bad stick of ram


Ok, so, look in bios update notes, if there is one, and if that is inconclusive compare the recommended clock of the memory to the clock set in the bios? i have asuses v bios.


yep. and be sure your BIOS is not attempting to overclock your ram by tuning. Or just assume you have found a bad stick and replace it
December 26, 2013 11:21:37 AM

i do belie i have fixed the issue, i had the bios in "asus preformance mode" was applying a "slight" over clock according to asus, but it no longer crashes with rendering and that did it EVERY time, reliably 100 of the time, now when i do it, it finishes the render perfectly. Big thanks to johnbl for reminding me to look for this.

cheers.


(gona do some prime 95 just to check)
!