I've been having issues with my computer freezing when playing high performance games. During the freeze the sound will cut off and all controls will be unresponsive. The only way to fix it is to do a hard reboot. Prior to the freeze there will be intermittent sound stuttering and slight freezes. Other issues are monitors listing no display while the sound works in the background, again during high performance games. This issue seems like it happens at random.

I'm not sure how to diagnose what the problem is but I am assuming it might have something to do with my PSU not having enough power.

Thanks.

Have you done any BIOS updates lately? The first round of meltdown and spectre patches were causing freezes.

Looks like my BIOS are a little out of date. I'll try that, thanks.

1) Patches are definitely a possibiliy,
2) but I suspect it's the hardware. Your cpu is old, so is likely your mobo and RAM. Maybe your PSU can't deliver enough power anymore
3) also while temps in speccy show fine, they could go critical during load.

3) is easiest to check, so do that first:
https://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.34.zip

Extract, run and check if your temps for cpu, mobo and gpu show sane values (sometimes the sensors are not read correctly).
Then run a game. Ideally have hwmonitor open on the 2nd monitor so you can view values live. If not, tab out from time to time to see how temps change.

for Mobo/system anything below 60°C is safe
for CPU anything below 80°C is safe
for GPU anything below 85°C is safe

See if crashes and hangs correlate with exceeding safe temps, and if yes check cooling. Often it's just stuck or broken fans or dust buildup.

You can also check voltages with hwmonitor, to see if voltages noticeably drop during load (PSU problem). I can't give good guidelines which values are still acceptable, so feel free to post your findings so we can discuss them.

2) can be tricky to determine. Basically you need to remove / replace a component and see if that fixes your problem.
First you should remove formerly stable overclocks (if any) and/or underclock RAM and reduce power target for your GPU.
The former finds problems with RAM or mobo components, the latter an aged PSU.

Next you can do a visual check of your mobo. If there are any blown capacitors they are not performing perfectly anymore and thus probably failing during demanding states like gaming.
BLOWN = bad

If those steps turn up nothing useful start swapping components.

Removing RAM is usually easy. You don't need full RAM to run stuff. Try that first. It's also easiest to replace. You can also run a memtest boot cd to check if any errors turn up.
Ask a friend if you can swap their GPU with yours for testing or abuse a friendly return policy on a new GPU.
Do the same for PSU.

If you managed to swap all above parts and you still have crashes, you can try a complete reinstall of your OS. Otherwise it's time for a new system (mobo, CPU, RAM). If you really can't afford a new system there could be ways around it. Just ask if that's really necessary.

1) I have no idea what to do if it's the patches. I made sure to wait until things have calmed until I apply them, so maybe a system restore can help?

EDIT:
DON'T apply BIOS updates until you have figured out your problem. An old BIOS is never a concern for system stability but could introduce new problems, complicating the search for the culprit.

Thanks for the detailed write up. I'll test those things tonight. In the meantime here's a screenshot I took earlier while the computer was running a game. I'll take a new one with that other program.



I've recently replaced the cooling gel and got the new GPU. Although I suppose it could be defective.

the advantage of hwmonitor (besides also giving voltages) is that it keeps track of maximum temp, which you might miss when just looking at the temps intermittently. The temps from your screenshot are okay though.

Yeah, I should of been more clear. I meant like.. If you did a BIOS update recently it wouldn't hurt to check again as some of the recent updates had flaws. New versions are being pushed out now. You coukd download OCCTP (i think) it stress test your CPU andmonitors temps. If your GPU is idling at 71 that is a bit high depending on where it is.

it's at 38 when idling (see first image). The last one is during gaming.

Okay here's some of the info from HWMonitor and other stuff. The freezing still seems really random and happens more often now. One thing I forgot to mention is that my screens would go blank and afterwards it says my gpu crashed and recovered. It happens sometimes but doesn't occur during freezing.

This is what windows said after the monitors had no signal. It restarted automatically and provided this report.


Here is what the event viewer says after the gpu recovers.



Attached are 2 pics and the HWMonitor File. One from HWMonitor during a freeze, another from after a restart. The HWMonitor txt is from gaming where I didn't experience any issues. I'll work on a visual check and ram/memtest next.

Is there a program that will give a better detailed log than event viewer?

Attachments:

Post has been edited 2 time(s), last time on Mar 1 2018, 7:59 am by Sie_Sayoka.

More info on the crahes / recoveries / blue screens / check disks*? No. It's already pretty detailed.
* a full report of the checkdisk is put into event viewer after boot:


The temps seem more than fine on the hwmonitor screenshot. I'd like to the see the values for the gpu too though. And maybe another screenshot which is fairly close to before a crash.

The voltages look concerning, but as I said, that could be a wrong interpretation of the sensor input. At least most of them are stable which is also an important metric. Again, the gpu voltages during load and the idle values would help to get a better picture.

To get all of the interesting data on one screenshot collapse the uninteresting ones: Fans, Fans PWM, Powers, Clocks, Utilization, HDDs

For comparison of the odd voltage readings, get a screenshot from speccy in the motherboard section.
Another thing that can't hurt to check in Speccy is whether the S.M.A.R.T status (last value) is "Good" for each Hard Drive.

Post has been edited 2 time(s), last time on Mar 1 2018, 10:48 pm by NudeRaider.

Possibly a bad factory overclock. Try turning on Debug Mode in the Nvidia Control Panel and see if anything changes.

Hmm, I'm having a similar issue on my setup (I suspect overheating but need to confirm - how long does a water-cooling system normally last?). Hope you get to the bottom of this.