Ryzen 5800x turning off randomly either when idle or gaming

[MarkA]

I guess the motherboard is over 30-days since you bought it? If it isn't, you should be able to claim a full refund inc shipping costs, or request a repair/replacement of the faulty product.

Sadly, bought it in May. Only started hanging this last month IIRC.

 

[SeeNoWeevil]

You don't run NZXT CAM by any chance? There's a long running issue of hard lockups with it running.

 

[MarkA]

You don't run NZXT CAM by any chance? There's a long running issue of hard lockups with it running.

I do not. Thanks for the suggestion though!

 

[MarkA]

New board in, same hang. LOL
(New board = Asus ROG Strix x570-f gaming)
Re-testing memory, swapping them out with a mate later.

 

[Tute]

Try disabling CPU C-States in the BIOS.

Also if you can find/access it, disable Memory "Power Down Mode".

I think i've been fighting exactly this "bug" for the past 2-3 months, at first I thought it was the house electrics playing up but now from googling it looks like Ryzen doesn't give itself enough voltage when it's idling and it's (in my case) tripping the PSU relay and making it reboot.

My setup: Asus Crosshair VIII Impact, 3700X, 8Pack Ripped 3600MHz DDR4.

 

[henson0115]

try running your memory at stock 2133 and testing for crashing, then go from there.

 

[Emlyn_Dewar]

We never want to blame the CPU, but it sounds like it's a duffer.
AMD were pretty awesome when I RMAd a CPU, although that was a Phenom II...
Worth raising a ticket with them, and saying evrrthing that's been tried.

 

[MarkA]

Try disabling CPU C-States in the BIOS.

Trying this now.

try running your memory at stock 2133 and testing for crashing, then go from there.

Got my friends memory in, that had crashes too, both in 2133 mode.

We never want to blame the CPU, but it sounds like it's a duffer.
AMD were pretty awesome when I RMAd a CPU, although that was a Phenom II...
Worth raising a ticket with them, and saying evrrthing that's been tried.

It does look like it could be CPU - I'll check with the place I bought, to see what they are willing to do for me, especially since i've already purchased a new PSU and board to troubleshoot all this.

 

[Joxeon]

It does look like it could be CPU - I'll check with the place I bought, to see what they are willing to do for me, especially since i've already purchased a new PSU and board to troubleshoot all this.

Your best going to AMD directly as not only will it be faster to get a replacement since if you go to the retailer than they will only send the CPU off to AMD and then on receiving a replacement send it back to you so adding a extra delay but often the retailer will test the CPU first by running prime or something similar and since the error won't show up in heavy stress testing then they may even just return the CPU back to you saying that there isn't anything wrong with it.

 

[Tute]

Trying this now.

@MarkA had any crashes so far?

I've been reading up on this for most of the evening and here's what i've learnt. The Ryzen CPUs have a power state called C6 which allows them to disable individual cores when not in use and drop the CPU voltage all the way down as far as required, even to zero (!) if necessary (what's the point, it's basically off?).

When the CPU drops into this state it seems that it's not applying enough voltage to keep itself alive, which means there's either a bug in the CPU itself or it's not stable at that certain clock, with those cores disabled, at that voltage, which means a binning issue.

Either way, i've found reports of people with first gen Ryzen processors getting the same issues back in 2018. And there are lots of them. So AMD have completely dropped the ball here - this should have been addressed at BIOS level years ago. The fact that it's persisted through 1000, 2000, 3000, and now 5000 series is absolutely appalling.

The suggestions vary wildly but most get success from disabling C-States, which stops the CPU going into C6. But it also stops it leaving C0, which means you will never drop below stock voltage. This has heat and power draw implications.

Hope that helps.

 

[Tetras]

Your best going to AMD directly as not only will it be faster to get a replacement since if you go to the retailer than they will only send the CPU off to AMD and then on receiving a replacement send it back to you so adding a extra delay but often the retailer will test the CPU first by running prime or something similar and since the error won't show up in heavy stress testing then they may even just return the CPU back to you saying that there isn't anything wrong with it.

With CPUs that have freezing / idle issues (or anything not super obvious) I'd definitely go to AMD, you'll almost certainly just get it sent back if the retailer tests it.

 

[MarkA]

@MarkA had any crashes so far?

I've been reading up on this for most of the evening and here's what i've learnt. The Ryzen CPUs have a power state called C6 which allows them to disable individual cores when not in use and drop the CPU voltage all the way down as far as required, even to zero (!) if necessary (what's the point, it's basically off?).

When the CPU drops into this state it seems that it's not applying enough voltage to keep itself alive, which means there's either a bug in the CPU itself or it's not stable at that certain clock, with those cores disabled, at that voltage, which means a binning issue.

Either way, i've found reports of people with first gen Ryzen processors getting the same issues back in 2018. And there are lots of them. So AMD have completely dropped the ball here - this should have been addressed at BIOS level years ago. The fact that it's persisted through 1000, 2000, 3000, and now 5000 series is absolutely appalling.

The suggestions vary wildly but most get success from disabling C-States, which stops the CPU going into C6. But it also stops it leaving C0, which means you will never drop below stock voltage. This has heat and power draw implications.

Hope that helps.

I left it idling over night after disabling C-states.
No crashes.

[image below is a photo with loads of reflection, on my raspberry pi-lcd, sorry)


You can still see the vcore is still fluctuating, which is what I expect. I just don't expect it to start dropping off unused cores now? I'm happy with that, as long as temps stay low, especially at idle. I'd be happy if this fixes it. I may test out with AMD, to see how long their RMA process takes..

 

[henson0115]

sounds like a duff cpu based on everything you've said and tried. rma it. you shouldn't have to disable c states like that for it to work. theres clearly an issue with the voltage regulator.

 

[MarkA]

sounds like a duff cpu based on everything you've said and tried. rma it. you shouldn't have to disable c states like that for it to work. theres clearly an issue with the voltage regulator.

Thanks, I've logged an RMA. Waiting to hear on estimated swap-out end-to-end period.

 

[humbug]

Good Grief... what a mess, how are you still so calm? i would be climbing the walls at this point.

 

[henson0115]

Thanks, I've logged an RMA. Waiting to hear on estimated swap-out end-to-end period.

hopefully you'll get a quick turnaround

 

[MarkA]

Good Grief... what a mess, how are you still so calm? i would be climbing the walls at this point.

I dunno really, I just try not to let it get to me and jump through the hoops that are throw at me.

hopefully you'll get a quick turnaround

Cheers!

 

[MarkA]

Your return AMD product has successfully passed the inspection and your
replacement product is now approved.

Woop woop! Hope it's not long until they give me a new one.

 

[Nelly]

Woop woop! Hope it's not long until they give me a new one.

Did you get it sorted out? I noticed its been over a month.

 

[MarkA]

Sorry. Yes, new cpu. No more issues whatsoever.
Glad i finally got there!
Interesting one though, with about a week without a gaming PC, I’ve actually found other hobbies and have lost the will to play games! I don’t miss it at all.

Happy with AMDs turnaround though!