Nailing down system instability

brafu

Honorable
Jan 24, 2013
75
0
10,640
Ok, I have a new build that's giving me a lot of headaches and I just can't seem to nail the issue down. At this point I'm just tired of trying to resolve this because I lack the knowledge and experience to properly figure this out.

How do I go about testing each piece of hardware in order to figure out where my issue resides?

I have already run OCCT, Prime95, and Intel's extreme tuning utility to stress test various elements of my PC, but all passed. Also ran Windows Memory testing utility and the RAM passed.

If more detailed info is desired I'll post it, but it's a long story lol
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


I have the latest drivers and a hard drive issue seems remote considering the issue.

The issue?

My PC is freezing during games, requiring a hard reboot. During normal use (ie. anything non-gaming) I have no problem at all. I have gone through a lot of testing (especially for heat issues) and at this point I'm down to wanting to test each piece of hardware to make sure they are all fine and that it's not a faulty component (the PC is about 2 weeks old).
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
i5-3570k (stock fan, no OC... in fact I even underclocked and still had issues)
Intel H 4000
ASRock Z75 Pro3
Crucial Ballistix Tactical Tracer 4GB (2 x 2GB)
CORSAIR CX430M 430W ATX12V v2.3 80 PLUS BRONZE
old 80gb drive I had in an 07 desktop (plan on replacing it soon)

From my understanding my PSU should have absolutely no issues powering all of that (which is the extent of what I have atm), so long as it's not faulty. The drive might be an issue, but I would expect the freezing issue to also effect normal operations, or other programs.

The only way I was able to reproduce the freezing I experienced in multiple games (ME1/ME2/ME3/SC2) was to run OCCT (GPU test) at the same time Intel's CPU stress testing software was running. The combination of CPU/iGPU load was enough to cause a freeze. Heat would be my first guess (and it was), but I have watched it freeze up anywhere from 60-75c more than once during such a test... yet when I ran Prime95 for almost 50 min the average temp was 70-75c (spiked up to 80c a few times) and there was not so much as a hiccup.

I'm aware that's not even close to ideal temp, and I plan on fixing the problem soon. But from all my research that temp, while not ideal, shouldn't be enough to cause the CPU to shutdown to protect itself. FYI, idle and normal use is around 30-35c.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


Then that would have to mean it's a faulty PSU, right? Because there's no way my system should be drawing enough power to tax that PSU with just the MB, CPU, HDD, and RAM. I don't even have a GPU yet.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
I tested my drive and it came up ok, and I also got that Whocrashed program and there wasn't any dumpfile for it to read (made sure it was even enabled, and it was).

Very frustrating when my brand new build is unusable for it's intended purpose lol
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


It was a fresh install of Win 8 Pro, full format. I have even refreshed the OS and uninstalled the video drivers to no avail.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


I don't believe it'll be possible to borrow another PSU (or any other component), but I'll see if the PSU from the old desktop will work for a test (probably too old though).
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
I checked the other PSU and it wont work for my board. First it's a 230w unit, and secondly it doesn't seem to have the right connectors for the job. None of that is too surprising considering its an old dell (dimension 3100 I believe).
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
I just had another freeze and reboot (rebooted on its own this time) while playing ME1 in window mode while I watched the various sensors. The CPU and GPU were drawing around 30W's or so combined (if the sensors are correct), and the temp was around 50c. The GPU wasn't even really being used at that particular moment (at-least not much) and yet I still had a freeze.

I wonder if throwing the entire PC in the trash and starting over would be an overreaction? =P
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


I ran memtest overnight for a total of about 8 passes and no errors were detected (had both sticks in).

So far today I have been busy with putting in the 2.5 HDD from a netbook I have (has a broken screen, but otherwise fine) and removed the old 80gb HDD. I'm in the process of installing some must have app's and then I'll install a game and see what happens.

Right now I'm leaning towards it being a HDD issue because yesterday the computer froze up when I was opening steam, so that expanded the possible causes. Hopefully this other HDD will resolve the issue.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
Well, crapped out on it being a HDD issue since I got yet another freeze while playing a game. I'm going to retest my RAM and then stress test the CPU and GPU separately, and then together. Figure I'll spend the day letting those run and see what happens.

EDIT::

The more I think about it the more I think that it really might be my PSU since it seems like the only thing that makes sense now, at-least so far. Without having another PSU to use, or any tools for testing voltage and such, what's the best way to tell if it's a faulty PSU? I'm going to attempt to draw as much power as I can and see what happens, but if anyone has any advice I would appreciate it greatly.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640
Ok, I'm pretty sure it's a faulty PSU because the only way I can reproduce the freeze is when I put the CPU and GPU underload at the same time. I can stress one or the other (or one of those and RAM) and I'm fine, but once I combine the two I get a freeze in under 30 sec most of the time.

I have been using OCCT's PSU test to help me figure this out, and so far a faulty PSU seems the most logical culprit. So I'm going to try and get a RMA from Corsair and get a replacement.

Thanks for the help and advice.
 
You wrote "The only way I was able to reproduce the freezing I experienced in multiple games (ME1/ME2/ME3/SC2) was to run OCCT (GPU test) at the same time Intel's CPU stress testing software was running. The combination of CPU/iGPU load was enough to cause a freeze." You're now convinced the PSU is faulty, but the problem probably is the IGPU that gets too hot. Using a better cooler or a graphics card would most likely resolve your issue.
 

brafu

Honorable
Jan 24, 2013
75
0
10,640


I thought heat was the issue at first, but after monitoring temps I'm convinced that's not the issue. Most of the freezes I have reproduced have been when my CPU has been around 60-70c (with and without the Turbo OC), so I really don't think it's a heat issue at this point. Plus I have stress tested just the GPU a good deal and it's never froze up during those tests.
 
How are you monitoring the IGPU temperature to make sure it's fine when both the CPU and IGPU are under stress? I doubt a 430W PSU would have issues providing power to that system, but it's possible if there's an issue with it. What's your next step if replacing the PSU doesn't help? Will you RMA the CPU or install a better cooler or a video card?
 

moulderhere

Distinguished
Sometimes I like to get a second opinion on temps on GPU. So I use my Canadian tire laser pointer thermometer and point it at the heatsink. Thus comparing gpu software temp with heatsink temp.
Usually are close in values.