Question System halts AFTER gaming ?

[Thrashermosher]

Hello everyone, I've just built a new rig after nearly 10 years without hardware upgrade and face really strange issue. I've been a huge PC enthusiast for years but had to take a break for last 5-6 years because of real life load. Anyways, here's my current setup:

CPU: i9 12900k cooled by noctua NH-D15
MB: Asus prime Z690 D4 - P
RAM: 32 Gb (2x16gb) DDR4 Kingston fury 3200mhz
GPU: Zotac RTX3080 trinity
SSD: m2 nvme wd black 1tb

Everything powered up with Kingston 750w psu which looks really solid and packed inside fractal design vector rs case.
OS is windows 10 pro as I have license for it and not willing to spend on win11 now

First of all, performance is a true killer, BF2042 runs smooth on 2k@144hz with everything maxed out, others games also cool tho Im not playing much anything else. Can play for several hours without any issues, CPU temp won't go above 75 C which is okay under heavy load and GPU stays at around 80-85.

Once I'm done playing and quiting to desktop PC works for next 15-30 seconds and than freezes completely, mouse coursor stucks and keyboard gives no respond, so I have to press power button to restart it. After that everything works great again. I thought its some battlefield 2042 issue but today this happened after playing civilization 6 which requires far less system resources.

So far I tried: bios upgrade to latest build, sfc /scannow, ddu gpu driver reinstall - nothing helps.
Hoping for community help and experience here, thanks!

 

[DRagor]

Welcome to the forum Thrashermosher.
Not having answers for you, just a couple of questions to helps us understand problem more.
1> Does it happen always when you finish game? If not, how often? Does it seem to happen more or less often for other games?
2> You say game works fine for hours. Did you try to invoke those freezes with short (up to 5 mins) game sessions?
3> Does it happen when not gaming? If you didn't do this yet, run such a test: let the PC running whole night under normal load (say, showing a looped video). Will it freeze?
4> Run a Task Manager (process tab) on second screen when you shut the game. Make a picture of it with phone once it freezes. Repeat it several times. Compare the pictures and look for anything unusual.
5> 15-30 seconds is about the time it takes for GPU to cool down after closing the game (CPU cools much faster). Take a closer look at GPU temps when freeze happens. Maybe it always freezes at the same temp?
That would all for now off the top of my head. Your problem is sure unusual. It won't be easy to pinpoint its source.

 

[DRagor]

Oh, one more thing: check the event viewer to see if those freezes left any errors/warnings there (like the dreaded "driver stopped responding" error).

 

[Thrashermosher]

1. It happens after heavy load, I'd say 8 times out of 10 playing sessions. With some older games it doesn't happen, e.g. fallout76 or dota2. But im not playing em, just installed to try out
2. With short sessions freeze doesn't happen. Only when system works under load for some time
3. No, with regular load chrome/video/vs studio everything is smooth
4. Ill give it a try as I use tv as second screen, thanks
5. Overall system cools down fast. Yes, I have some ideas about it as I put 3 enclosure coolers into single pwm controller and after that put pwn cable into cha_fan1 pins on motherboard. HWinfo shows over 65000 rpm+ while idling, which is definitely a false reading. I will try to disconnect it but without case coolers but temperatures will be pretty high without them.

@DRagor

 

[Thrashermosher]

Oh, one more thing: check the event viewer to see if those freezes left any errors/warnings there (like the dreaded "driver stopped responding" error).

It sends a kernel 41 error which is pretty typo and can explain a wide variety of issues

 

[DRagor]

Ok, you know where it leads? Normal loads and short sessions of heavy loads don't cause freeze - because they don't raise temps high enough. So this does looks like a overheating problem, although definitely not typical.

I will try to disconnect it but without case coolers but temperatures will be pretty high without them.

Don't do it. Going up with temps is definitely not the way to solve the problem.
I'll repeat myself: check GPU temps at freezes and see if it happens at same temp. Also, check motherboard temps. Motherboard cools down slowest of all the parts (mostly because most setups don't have fans set up to cool board directly).

 

[Thrashermosher]

Ok, you know where it leads? Normal loads and short sessions of heavy loads don't cause freeze - because they don't raise temps high enough. So this does looks like a overheating problem, although definitely not typical.

Don't do it. Going up with temps is definitely not the way to solve the problem.
I'll repeat myself: check GPU temps at freezes and see if it happens at same temp. Also, check motherboard temps. Motherboard cools down slowest of all the parts (mostly because most setups don't have fans set up to cool board directly).

Yeah that's right, but actually problem occurs when temperature goes down, for everything inside pc case. At heavy load, during game session temp may go high, but yet not crossing critical values, 75 c for cpu and 85 for gpu is maximum I could achieve with some hours of non-stop bf2042 playing. But once I stop, watching temps goes down (HWmonitor is what Im using), temp drops to average 45-50 C for cpu and 50-55 for gpu. All motherboard sensor return regular temps aswell. It feels like system goes crazy because of temperature goes down way too fast lol

 

[Thrashermosher]

There's also something I forgot to mention - DDR temp goes completely wild to 80 degrees C, which looks crazy for me. Also I can't make system start with XMP profile, only with manual or auto settings. This is pretty confusing but I didn't do any deeper research as I plan to migrate to DDR5 by end of 2022. Could ddr be an issue

 

[DRagor]

Whether the problem is for temps too high or too low, you need to find out which component is doing trouble first. It does not seem to be CPU, the rest are suspects.
I don't think I ever seen a board that reports RAM temps - maybe that's new feature in Alder Lake boards. In that case, does the sensor reports RAM temp or motherboard temp near the RAM? You may want to check the temp section in BIOS to find out (or perhaps board manual). Now not able to use XMP is not some something unusual, esp, with new board like yours. If the problem was with RAM settings it would be more likely to cause problems during playing. But then again RAM can overheat, as well as motherboard VRMs powering it, although this takes setting wrong voltage in most cases (and Auto setting could do that).

 

[Mariusglock]

Hey Thrashermosher, i have the same motherboard, unfortunetelly this MOBO has issues with XMP
Before BIOS update i couldnt run XMP, nor manually set it. After Updating BIOS i had Bluescreens with XMP.
What did the trick is changing RAM slots, so now i use A2 and B2 slots, and it works perfectlly
So i would suggest trying to change to these slots, or trying 1 RAM module at the time
And are you on BIOS ver. 0605 now?

 

[Thrashermosher]

Hey Thrashermosher, i have the same motherboard, unfortunetelly this MOBO has issues with XMP
Before BIOS update i couldnt run XMP, nor manually set it. After Updating BIOS i had Bluescreens with XMP.
What did the trick is changing RAM slots, so now i use B1 and B2 slots, and it works perfectlly
So i would suggest trying to change to these slots, or trying 1 RAM module at the time
And are you on BIOS ver. 0605 now?

Yo, hello! Yes, bios is 0605, I updated it with official asus image this morning. B1 and B2 will be just under my noctua cooler, will be hard to reach them, but ill give it a try. There's something wrong with it I believe because kingston fury modules should be fine, been using kingstons for ages with different rigs. Should we bomb asus support with that?
@DRagor I hear you, HWmonitor reports ram temp, not sure if its z690 feature or something. Also not sure 80 C degrees is ok for RAM. There's something hanging my system ones load goes off and temperature and hz dropping back to idle. I believe its motherboard issue afterall but its impossible for me to try with another one, unless asus or another vendor sends me one for tests 😆 lol

btw I ran a memtest86 overnight yesterday and it reported 0 errors with 4 passes

 

[--SID--]

Kingston 750w psu which looks really solid

Never heard of. How it looks outside doesn't matter for the functionality. It has to be solid inside. Do you have a picture of the label?

 

[Thrashermosher]

Never heard of. How it looks outside doesn't matter for the functionality. It has to be solid inside. Do you have a picture of the label?

I'm sorry, my bad, its Corsair, not Kingston. Corsair RM750X one

 

[DRagor]

@DRagor I hear you, HWmonitor reports ram temp, not sure if its z690 feature or something. Also not sure 80 C degrees is ok for RAM. There's something hanging my system ones load goes off and temperature and hz dropping back to idle. I believe its motherboard issue afterall but its impossible for me to try with another one, unless asus or another vendor sends me one for tests 😆 lol

btw I ran a memtest86 overnight yesterday and it reported 0 errors with 4 passes

Like I said never seen board report RAM temp before so no idea is it good. But I somewhat doubt it's RAM temp - you would need temp sensor on RAM stick to do that.
Might as well be a problem with board, after all it's new build so infancy problems might happen. When first gen Zen boards hit the market it was almost impossible to find RAM that was working with them.
As for memtest result, I would not expect anything else. After all your system works fine under load

 

[Thrashermosher]

Okay little update;
I managed to run XMP profile, for this I put ddr modules in double channel slots A2 and B2;
Went to bios and picked xmp #2 which is DDR3000, then I manually put ram voltage to 1.35000 v, trick worked and windows loaded without a prob. I did a test match with battlefield with everything maxed out, rtx off and DLSS set to auto, first 10 minutes I had spikes but after that they were gone and went to stable 120 fps with jumps to 140 and drops to 90.
I used OCCT software to check temps and power which told me my CPU was max. 67 degrees, GPU 70 (however gpu vram still hits 80)
CPU+GPU power consumption is about 300w total which suits me just fine. After nearly a 25 minutes match I had no freezes and PC works as intended after cooldown.
Ill keep watching as I believe problem may occur again. Tomorrow Im planning to migrate to win11, looking for place to buy legit key atm.

Also, I haven't connected the 4 pin power cable to motherboard, I simply don't have a 4 pin cable with PSU. Not sure if thats okay or not, mobo book says nothing

 

[DRagor]

Also, I haven't connected the 4 pin power cable to motherboard, I simply don't have a 4 pin cable with PSU. Not sure if thats okay or not, mobo book says nothing

I suppose you did connect 8-pin cable to motherboard though, otherwise your PC would not start. Adding 4-pin to that only changes how much power CPU can use. I strongly suggest you connect it - and yes you do have it. Your PSU should have 2 8-pin cables called EPS12V. The 8-pin connector on those cables split in half to two 4-pin connectors that you need. It might be a little tricky if you never did it before so maybe watch some youtube video.

 

[Thrashermosher]

I suppose you did connect 8-pin cable to motherboard though, otherwise your PC would not start. Adding 4-pin to that only changes how much power CPU can use. I strongly suggest you connect it - and yes you do have it. Your PSU should have 2 8-pin cables called EPS12V. The 8-pin connector on those cables split in half to two 4-pin connectors that you need. It might be a little tricky if you never did it before so maybe watch some youtube video.

Well mate, that's what I read too but yes, Ive connected a 8 pin one so it starts. ill take a closer look.
I did a second test with 40+ mins game - no crash.
Did a unigine benchmark on 1080p extreme and it went flawless (results aren't superb however, only 10508 points)

Added: I just checked again and yes, one 8 pin can be used as two 4 pin, damn, shame on me rofl. Was thinking of soldering one myself from old PSU lol

 

[Thrashermosher]

Okay, I'm here with little update;
I finished working with PC last night and turned it off as usual,
This morning I decided to read some news and browse web, but it failed to start. I've tried to turn it on and off several times with power button but nothing worked, so I decided to put a 4-pin additional power cable asap.
After that it booted normally and I managed to enable XMP1 profile without problem with maxed DDR3200 setup.
So far I played a test game for 25 minutes without crash or something, temperatures rised a little bit but I believe thats because of XMP profile takes more power/performance now.
However I see no way to check if newly applied 4-pin power cable actually works because bios won't show me detailed power information.
Thus, I got a +150 points growth in unigine benchmark.
I will keep you updated
Here's HWmonitor voltage readings:

https://ibb.co/SXmqhSw

IMG-20211219-135849 hosted at ImgBB

Image IMG-20211219-135849 hosted in ImgBB

ibb.co

OFFSET voltages used for overclocking I believe, since Im not doing any I think its okay to have them at 0?

 

[Thrashermosher]

Welp now I can't set XMP profiles again, PC won't start unless I put everything to "auto".. guess we should wait for bios update from asus
Edit: I've submitted a ticket to asus support as they have my ram in capability list (exact #) I honestly believe there's nothing I can do about it, probably @Mariusglock will have some updates regarding this issue too

 

[Thrashermosher]

A little update, I got an answer from Asus support where they suggest to install only one ram module into A2 slot and try to start system by entering xmp values manually. If that won't help they suggest to contact RMA and return board. However, if that helps they don't say what should I do with my second 16gb module.
Make ram is in asus supported hardware list too

 

[Mariusglock]

Also, I haven't connected the 4 pin power cable to motherboard, I simply don't have a 4 pin cable with PSU. Not sure if thats okay or not, mobo book says nothing

This mobo has 8pin and 4 pin connection (so 2 cables) for the CPU so connect them all, dont leave anything out

Anyways about the RAM, as I sayd this mobo has some issues with XMP or manual settings.
I got it working after BIOS update and putting it in A2 - B2 slots.
But i guess your ram modules that you have, does not satisfy the board.
Just return this board, a lot of people have issues with RAM on this board, and theres no assurance when or if will they fix this with a new BIOS update.
return it and get your self a new board

Or you could try once again, reconnecting everything as it should be ( since you mentioned you were running your system without one of the CPU cables), then reset you bios to default, make sure your on the latest bios. And insert your rams to A2-B2 (2nd-4th slots).

 

[DRagor]

A little update, I got an answer from Asus support where they suggest to install only one ram module into A2 slot and try to start system by entering xmp values manually. If that won't help they suggest to contact RMA and return board. However, if that helps they don't say what should I do with my second 16gb module.
Make ram is in asus supported hardware list too

Worth trying it. If that works, there's good chance it will still work when you add second RAM stick.

 

[Thrashermosher]

I will give it a try today or tomorrow, bit away from rig atm. Its inside A2 and B2 now and I already connected 4-pin power source too. So my only hope is making it work with one module and than trying to add second one. Otherwise Ill stay with it till end of 2022 when Im planning to migrate on ddr5

 

[Thrashermosher]

Okay so I got back home today and tried asus support recommendations. PC works like a charm, no crashes, boots up okay but I have to use single 16gb module with A2 slot. Also, I had to put custom timings, ill post it here so it may help someone else:
DRAM CAS # Latency -> [16]

DRAM RAS # PRE Time -> [20]

DRAM RAS # ACT Time -> [32]

DDR set to DDR3200 manually and voltage to 1.3500V aswell.

 

[Thrashermosher]

Failed to boot up again this morning, but after restart it booted ok without reporting a POST error message. I've added second module into B2 to its dual channel with asus suggested settings. Now looking forward to benchmark my CPU as I have some doubts.
Im using battlefields built-in FPS counter, it shows FPS rate for both cpu and gpu. Game works okay with everything maxed out, dlss and rtx on, but with rtx on game reports fps drops on CPU side, it goes to 45-50 sometimes while gpu holds stable 70. Game feels fine still and to be honest I don't like RTX effects, so i usually turn it off, but this cpu fps drops made me cautious
Edit: with two ram modules issue is back, PC won't start and after several restarts reports POST error and drops RAM settings to default ddr2333. Pretty much sad and disappointed with asus