Random PC freeze while gaming. PSU the victim?

[PCI_Slot]

Please refer to the PC specs on my profile.

I cant recall when this started, but its been going on for a while (1.5years?).
Everytime it happens, I noticed it is on certain games which I would say is quite graphically intensive. When my PC hangs, I observed my GPU temp is at ard 75deg.

Weird thing is that it wont happen instantly, it will plateau at 75deg and the hang will come in randomly. I mean if its a heat issue, I'm pretty sure that I can reproduce the problem easily and it will straight away hang once it reaches 75deg or atleast hang after maybe 10secs at 75deg. Like for example, Overwatch is a game that have quite constant graphic activity (right?), my PC will hang at any time...

How the hang looks like:
Screen freeze with whatever is showing on the monitor, frozen as it is. No graphical artifacts.
As for sound, depends, sometimes it will repeat the last sound thats playing (irritating sound) or sometimes just silent (Seems like depends on the last milisecond of sound thats was being emmited)
Have to press and hold the power button to turn the PC off.

The diagnose:
Previously I was using a GTX 960 2GB. I managed to get my hands on a 2nd hand GTX 1070 8GB recently. Both setup face the same issue, so I kinda concluded that it is not a GPU issue. Unless I damn unlucky both cards are faulty.

My guess:
1) PSU: After giving it a thought, I think it could be a PSU issue? This system is already 3.5years old (which includes the PSU). Could it be that the CX600 is not supplying enough, or "clean" power? Unfortunately I am unable to diagnose this yet, because I do not have an extra newer/more powerful PSU ard.
I guessed this because, when on low or normal load, no problem whatsoever because PSU is able to supply enough power. But when on high load and PSU is already delivering the most, when there are sudden spikes for power, the PSU fails to deliver. Does that make sense?

2) RAM: Bad RAM? I have not really tested it by installing 1 RAM each time. Shall consider this. But ive done Memtests before, no issues. If RAM's the issue, my guess is that Memtest should have already picked it out?

3) CPU: Have not test anything, but im not inclined to test this coz the Temp are not that hot when the hang happens (I remember that it is not an alarming temperature value)

4) HDD: Hmmm dont think so. When I do heavy disk activity work, have no issues whatsoever.

5) Motherboard: Haha this one I think can only conclude when all the rest have no issue.

 

[Godeau]

Possibly the Ram, Psu and/or Motherboard.

Memtest doesn't always pick up issues within the first hour of testing, and sometimes doesn't pick up issues at all(even after the whole night). Since it's free, can't complain much.

Psu, gotta shell out money for a new one if you don't have a spare or a friend's to borrow from.

Mobo, same as psu.

 

[limcc]

Stress the system with both Prime95 + FurMark. It will load both CPU and GPU to max, good test for PSU or motherboard if they are able to supply stable voltage/current.

https://www.tomshardware.com/reviews/stress-test-cpu-pc-guide,5461-8.html

With CPU and GPU at max load, you might cause some damage to the sytem if the cooling is insufficient or below par hardware are used. The OCCT PSU test should be a safer choice, all in one app with temp and voltage monitoring--from onboard probe.

https://www.tomshardware.com/reviews/stress-test-cpu-pc-guide,5461-9.html

 

[5star_pundit]

Try play overwatch offline mode
I suspect your internet hang

 

[SkyShroud]

I would say is GPU issue. Overheating might not be the gpu processor itself but your gpu memory.

 

[Phen8210]

i had the same issue awhile back, end up is my motherboard problem.

i ramtest and it always froze after random duration, but since it ran off my usb stick i can confirm is not my ssd problem.

Also i changed my PSU awhile back so that is less likely the issue.

Means it's either ram, mobo, cpu.

I rma the ram, the problem still there.
i rma the mobo, mission accomplished.

 

[yusoffb01]

How can culprit become victim?

 

[PCI_Slot]

Try play overwatch offline mode
I suspect your internet hang

eeeyerr... baptiste #teamzen

I would say is GPU issue. Overheating might not be the gpu processor itself but your gpu memory.

Ouu. If true, means quite unlucky leh coz i tried 2 GPUs alr, both same issue...
*See latest update below*

i had the same issue awhile back, end up is my motherboard problem.

i ramtest and it always froze after random duration, but since it ran off my usb stick i can confirm is not my ssd problem.

Also i changed my PSU awhile back so that is less likely the issue.

Means it's either ram, mobo, cpu.

I rma the ram, the problem still there.
i rma the mobo, mission accomplished.

Ya lor... hmm alot to diagnose. I can see that at the end of the diagnose, I will end up buying & having parts for another new set of PC hahaha
Maybe its a sign to upgrade my whole pc lol. Anw at the moment, im singling out my RAMs (Its a 8GBx2 configuration) to see if its a RAM issue or not.
*See latest update below*

How can culprit become victim?

*Latest update*
Yesterday night, after my PC hang while playing game (as usual, high GPU temp)
I force restart the PC to get back into the game, but it hanged again while only afew mins on desktop with GPU temp only 42deg. Might be a one off thing... will need to monitor.

 

[grimfan]

Set your RAM to less aggressive timings to test.

 

[Phen8210]

think ram or mobo, go run ramtest overnight.

the trick is to replicate the problem, mind is cannot replicate with stress test and games.. but somehow i run ramtest it 100% will hang.

 

[Godeau]

*Latest update*
Yesterday night, after my PC hang while playing game (as usual, high GPU temp)
I force restart the PC to get back into the game, but it hanged again while only afew mins on desktop with GPU temp only 42deg. Might be a one off thing... will need to monitor.

Desktop hang with no artifacts, although actual cause is still unknown, high probability is ram or mobo. Does your cpu have igpu? Use that without gpu and monitor. Or you can try using passmark's burnintest(it's not free, but comes with i trial i believe) to stress test all components

 

[PCI_Slot]

*Spoiler: Still monitoring the fix*

Oh well well well... what do we have here...

I did memtest 3 times with 4 passes each (because I didnt pay for the pro version that unlocks unlimited passes)

BIOS
Vendor American Megatrends Inc.
Version P1.10
Release Date 07/16/2015
Baseboard
Manufacturer ASRock
Product Name B150M Pro4

1st test:
Failed on Test 13 [Hammer test]
Test 13 [Hammer test] 3/4 (75%) 1
2019-06-28 00:42:51 - [Data Error] Test: 13, CPU: 0, Address: 1C104BA40, Expected: 292D688A, Actual: 212D688A

2nd test:
all ok

3rd test:
all ok

I did a quick read on Test 13 errors, seems like its not a common error, but it happens sometimes coz of use of cheap/low quality or old parts, or interference. Im guessing ECC memory solves the problem? But not worth since i dont do server stuff or money making jobs on this PC. (correct me if im wrong to take this error lightly, and its infact a hardware failure). Furthermore, the 2nd & 3rd test reported no errors.

Anyway, after Memtest error came up, I read alittle bit about RAM incompatibility and thought its a smart idea to check if I need to update my BIOS.

Lo and Behold, looking thru Asrock website, seems like my BIOS is quite outdated. Furthermore, the revision notes actually has things like
"Improve DRAM compatibility"
"Improve PCIe, memory compatibility"
"Update Skylake CPU Microcode to revision C2"

Current: 1.10 (07/16/2015)
Latest: 7.30 (13/3/2018)

I'll be honest, ive always been the type to make sure BIOS is updated on my previous pcs/laptops. And furthermore, this PC has been quite constant in giving me problems, I'm not sure why I haven't updated the BIOS last year . Could be because I was too fixated with it being a GPU problem all this while i guess, so I was just waiting for the day I can swap my GPU out. Which I did 2 days ago, and found out the problem is still there.

Anyways, I'm still monitoring if the updated BIOS solves the problem. I THINK it might.

There is also tabs on the website for CPU & RAM compatibility for the mobo. I have not really scrutinized the list, but quick look, my RAM is not inside the list. TBH, I did not know compatibility list is also a thing for consumer Mobo, as I assume that most hardware should work. Is this newly implemented? Previously I've been my Asrock mobo website, but did not come across it.

 

[PCI_Slot]

Btw, I have done some the tests suggested by the forumers here, THANKS! Some of the test I have not heard of.

I've done Furmark & OCCT PSU test, I loaded the GPU and CPU for 30 mins and nothing happened. So I was starting to feel that those might not be the problem.

For OCCT PSU test, I was only able to see the voltage from the CPU & GPU. Am I also able to see how much the PSU is actually operating at? Im guessing I cant see that value because my PSU does not have the sensor. Just wondering.

 

[Phen8210]

if you get the error most likely really have error and not the ramtest bug. see if you can replicate the issue again.

Seem like its pointing to your ram, mine was motherboard i more suey



try intelburn test also, set to highest allowed memory see what happens

 

[binary_0011]

have you try resitting your cards?