Random crashes...absolutely fed up!

Grandmaster Fap! · 5 Mar 2021 at 18:40

System spec:
2700X not overclocked.
Asus x470 gaming F
Team extreme 4000 8pack ram (16 gig)
Gigabyte 6800XT master
Bequiet Straight power 850w (6800XT is on two separate rails, not one cable split)
Various ssd

Ok so i went from a 1070 GTX and 600w Coolermaster psu with zero issues to a 6800XT and BQ 850w and all seemed well. Then a few days in id get random black screen and have to reset, no rhyme or reason, never seems to happen in games, always in desktop doing nothing taxing. Set PCI-E to gen3 in bios....uninstalled all gigabyte and asus software and had two days of stability then just watching YT and nothing taxing, comp freezes.....im fed up of this. Only thing i havent done is win10 new install.

To be clear ive run Timespy and Heaven endlessly and they always pass, Run OCCT endlessly and it passes, this is so random and its driving me insane.

Edit: Event viewer always shows the same crap, Kernel power event 41...useless.

pp111 · 5 Mar 2021 at 19:14

Have you tried the RAM without XMP set?

pc-guy · 5 Mar 2021 at 19:29

2700x wont do much over 3200MHz on the ram. what settings have you got for ram> have you tested that for stability.

also try disable all the C-state setting in bios and set power plan to min 100% see if that improve the situation.

also update the AMD drivers and that should install a AMD specific power plan in windows.

change the idle power setting in Bios to the other available option.

do the above changes one at a time to see if you get any results otherwise you wont know what is at fault. however i would change the C-state and the min 100% CPU power at the same time as they are effectively linked setting.

Grandmaster Fap! · 5 Mar 2021 at 19:35

pp111 said:
Have you tried the RAM without XMP set?

Its not set as xmp would set the timings crazy loose, i just set it to 3400 with 15 timings and always been great.

pc-guy said:
2700x wont do much over 3200MHz on the ram. what settings have you got for ram> have you tested that for stability.

also try disable all the C-state setting in bios and set power plan to min 100% see if that improve the situation.

also update the AMD drivers and that should install a AMD specific power plan in windows.

change the idle power setting in Bios to the other available option.

do the above changes one at a time to see if you get any results otherwise you wont know what is at fault. however i would change the C-state and the min 100% CPU power at the same time as they are effectively linked setting.

Stability on ram tests seem great but havent checked for a while as ive never had this issue. As for the C state settings thats a bit above my pay grade so you may need to walk me through that mate, im used to intel and nvidia (just working) so all this is a wake up call

pc-guy · 5 Mar 2021 at 20:03

I think you need to dial back your ram speed to 3200MHz. Even 3200Mhz is considered a pretty decent overclock on zen+ chips. Test that setting with memtest86 to see if the ram is stable and error free. You ram no doubt will go into 4000MHz however the ram controller on the CPU (same for intel and amd for some time now) so it really depend on the controller’s ability to run ram at all those speeds.

I would check out the ram first. You have done your stress test for GPU and possibly CPU. To test CPU I normally would also use AIDA64 and test for system stability. I run that for at least 2hrs and if no issues the. Everything should be fine.

If you still get reboots then I will go into bios and switch off all the C state settings. C states should be found under CPU feature. You can also search for it in your bios setting.

Tetras · 5 Mar 2021 at 20:20

Does the PC 100% lock, or is it only the visual?

Does cycling the monitor inputs help?

Have any spare cables?

Grandmaster Fap! · 5 Mar 2021 at 20:30

Tetras said:
Does the PC 100% lock, or is it only the visual?

Does cycling the monitor inputs help?

Have any spare cables?

100%. I had braided cables but switched to original and still had the issue. Monitor is dead when it happens.

Tetras · 5 Mar 2021 at 20:32

Grandmaster Fap! said:
100%. I had braided cables but switched to original and still had the issue. Monitor is dead when it happens.

What about if you switch from e.g. hdmi to dp?

Grandmaster Fap! · 5 Mar 2021 at 20:33

Here's the weird thing, i can run OCCT or Heaven while surfing or even running a game like 2077 or Control and everything works fine, but randomly while just doing nothing but surfing or YT is locks....its bizarre!.

Grandmaster Fap! · 5 Mar 2021 at 20:34

Tetras said:
What about if you switch from e.g. hdmi to dp?

Im on DP at the moment, will do a sweep tomorrow and see if that helps.

philo-sofa · 6 Mar 2021 at 02:12

I’d give display driver uninstaller a go if you haven’t already, remove anything Intel, AMD or Nvidia and reinstall your AMD drivers.

https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

tres_kun · 6 Mar 2021 at 12:12

i had an issue that might be similar to yours
The system was 100% stable under load
But will crash randomly
With trial and error i concluded that the memory was running to hot at the settings i was running on them
Under full load the case fans provide enough airflow to cool the ram
But at idle the airflow was close to nothing making the ram overheat
By overheat i mean 50 degrees, stability is temperature dependent on my settings

Showboat · 6 Mar 2021 at 12:17

I found RAM issues when i tested each one individually.

Grandmaster Fap! · 6 Mar 2021 at 12:33

Thanks guys, interesting what you say about ram temps @tres_kun, until i put the 6800 and new psu in i had no probs but the 6800 pumps out a massive amount of heat, the back plate gets mental hot. I was going to watercool the card as the fans are obnoxious but stalled the idea when i started having these issues(thinking i might have to rma it)

adrichardson81 · 6 Mar 2021 at 17:06

Have you tried a hefty underclock on the memory? Try 3200 c16 at around 1.25/1.3v and see if that helps.

If you run with side panel off, do you still get the crashes?

Grandmaster Fap! · 6 Mar 2021 at 17:17

adrichardson81 said:
Have you tried a hefty underclock on the memory? Try 3200 c16 at around 1.25/1.3v and see if that helps.

If you run with side panel off, do you still get the crashes?

Actually trying this at the moment mate, 3200 at 1.2v, also reinstalled win10 as its been well over a year since i did a clean install so figured it might help

keyser · 6 Mar 2021 at 17:44

I was about to say sounds like reset bios to defaults and clean install time, then mem test for 24 hours, after that reseat all RAM and unplug/replug all cables. Overheating things should just throttle, not die like that, but yes make sure you have ok airflow in your case too.

Grandmaster Fap! · 6 Mar 2021 at 17:49

keyser said:
I was about to say sounds like reset bios to defaults and clean install time, then mem test for 24 hours, after that reseat all RAM and unplug/replug all cables. Overheating things should just throttle, not die like that, but yes make sure you have ok airflow in your case too.

Yeah i reset everything in bios except cpu volts as that thing likes to drink 1.4-5v in spikes which i hate. Win10 installed and just had a bash at CP2077, OCCT and watching YT, three hours in and all seems good(but then again last time i went nearly two days before it crashed again so its not saying much yet!)

Grandmaster Fap! · 6 Mar 2021 at 19:00

Yeah im pretty much done and rma'ing the card and psu, just sat in windows watching YT and sound freezes and then comp lock. Tried every stress test and diag prog and nothing comes back as wrong but this happens randomly so yeah, done

keyser · 6 Mar 2021 at 19:46

There is a recent BIOS update for you motherboard you could try that. And then clear CMOS again.

Version 5837
2021/02/09 14.77 MBytes
ROG STRIX X470-F GAMING BIOS 5837
"1． Update AMD AM4 AGESA V2 PI 1.2.0.0
2． Update AMD RAID UEFI driver
3． Improve system stability"