5800x stability issues

jim_britton · 22 Apr 2021 at 20:03

Ayup,

I'm after a bit of advice on a recent Ryzen 5800x purchase from Overclockers.

This replaced an existing 3700x in an MSI Pro-a X570 board with 16 gig of cas18 3600 ram (with an AMD-specific XMP profile), an 850watt PSU, Noctua DH15 cooler and RTX 3080 and various other bits and bobs including a Valve Index.

The system's been solid and stable for ages with the 3700x. Since putting the 5800x in I just can't get the thing to be stable when gaming. It's rock solid doing regular Windows stuff, it always boots and runs brilliantly. But in *some* games (specifically, Automobilista 2, Project Cars 2 and Assetto Corsa Competizione) it'll blue-screen after anything from 5 minutes to an hour. The error is logged in the event log as a "bugcheck 0x00000124" which I believe is a generic hardware fault error code. There are no other errors in the event log (no on-going WHEA errors or anything, just this BSOD).

I can run CPU benchmarks like Prime95 without errors for ages. The CPU will boost to around 4600 all core and 4820 single core when benching, with temps topping out at about 80 when running all core prime85 small tests. Temps in gaming are much lower (in the low 60s). But as I'm a VR player I never actually get to see the temps at the point when it BSODs because I have the goggles on. I also ran the Windows memory diagnostic (no errors).

I've been through some fairly standard trouble shooting steps - set the ram back to default, updated drivers (chipset & GPU), tried 3 different BIOSes (the latest 2 beta bioses and the last 'official' bios). I've also tried with PBO disabled in BIOS and lower TDC, EDC and PPT limits (80, 120 and 120 respectively) but nothing seems to make any difference - it'll run for a while in these games but eventually BSOD with this same bugcheck error.

Is there anything I've missed here? I'm not really sure what to do - should I start the RMA process? It feels like it may be a faulty chip, given that the 3700x has been running like a champ with not a single crash or hiccup for more than a year, but I appreciate that demonstrating that it's definitely the chip will be a challenge

philo-sofa · 22 Apr 2021 at 22:19

jim_britton said:
Ayup,

I'm after a bit of advice on a recent Ryzen 5800x purchase from Overclockers.

This replaced an existing 3700x in an MSI Pro-a X570 board with 16 gig of cas18 3600 ram (with an AMD-specific XMP profile), an 850watt PSU, Noctua DH15 cooler and RTX 3080 and various other bits and bobs including a Valve Index.

The system's been solid and stable for ages with the 3700x. Since putting the 5800x in I just can't get the thing to be stable when gaming. It's rock solid doing regular Windows stuff, it always boots and runs brilliantly. But in *some* games (specifically, Automobilista 2, Project Cars 2 and Assetto Corsa Competizione) it'll blue-screen after anything from 5 minutes to an hour. The error is logged in the event log as a "bugcheck 0x00000124" which I believe is a generic hardware fault error code. There are no other errors in the event log (no on-going WHEA errors or anything, just this BSOD).

I can run CPU benchmarks like Prime95 without errors for ages. The CPU will boost to around 4600 all core and 4820 single core when benching, with temps topping out at about 80 when running all core prime85 small tests. Temps in gaming are much lower (in the low 60s). But as I'm a VR player I never actually get to see the temps at the point when it BSODs because I have the goggles on. I also ran the Windows memory diagnostic (no errors).

I've been through some fairly standard trouble shooting steps - set the ram back to default, updated drivers (chipset & GPU), tried 3 different BIOSes (the latest 2 beta bioses and the last 'official' bios). I've also tried with PBO disabled in BIOS and lower TDC, EDC and PPT limits (80, 120 and 120 respectively) but nothing seems to make any difference - it'll run for a while in these games but eventually BSOD with this same bugcheck error.

Is there anything I've missed here? I'm not really sure what to do - should I start the RMA process? It feels like it may be a faulty chip, given that the 3700x has been running like a champ with not a single crash or hiccup for more than a year, but I appreciate that demonstrating that it's definitely the chip will be a challenge

You have tried almost everything. I’d suggest trying the memory at JDEC speeds (2133 MHz) as the memory controller is the weakest link with Zen 3; that might at least help isolate the problem.

Other than that it seems like you’ve tried everything and done it right.

t_turbo · 22 Apr 2021 at 22:21

What brand/model is the ram?
Try it without xmp and see if it still crashes, I had a bunch of trouble with patriot ram. The 5k series seem quite picky.

jim_britton · 22 Apr 2021 at 22:41

The ram is Corsair if I remember correctly. I've already tried it at it's default (non xmp) timings and it's made no difference. I'm out of ideas, certainly feels like I got a lemon

t_turbo · 23 Apr 2021 at 09:40

jim_britton said:
The ram is Corsair if I remember correctly. I've already tried it at it's default (non xmp) timings and it's made no difference. I'm out of ideas, certainly feels like I got a lemon

Sorry, I missed that in your original post.
Any chance you've bent a pin? Might be worth trying to reseat it after checking. If not rma it, you might have gotten unlucky.

*edit: was this a fresh windows install?

Szydero · 23 Apr 2021 at 09:47

What's your CPU core voltage? If you have it on a default then try to lower it by 0.1v

P.S I am running my 5800x at 4.6ghz/1.15v

jim_britton · 23 Apr 2021 at 10:03

it's not a fresh windows install, it's from the previous build with the 3700x (so about 18 months old) - it was a clean install when I bought this x570 mobo and the 3700x. As this is my main development box (as well as gaming rig) a reformat would be very painful.

I'm not undervolting or overclocking - it's all at stock. I could try voltage tweaks I guess, but I think it really should run stable at stock settings with a decent cooler and PSU.

I guess I could have bent a pin installing. I was properly careful and there were no bent pins when I removed it from the packaging and i didn't catch it on anything when I put it into the socket. Not saying it's impossible but i doubt it. I'll take a look when I re-seat it later today.

I've had Prime95 running all morning without errors, with and without AVX enabled in 'blend' mode and 'small fft' mode, no errors or crashes

StephenJ · 23 Apr 2021 at 10:43

If you still have the 3700x you could swap it back to make sure the issues are resolved and that it is 100% due to the new chip

welshrat · 23 Apr 2021 at 11:55

What's your power supply like ? as if its gaming that's causing it could it be lack of voltage somewhere when GPU is drawing more power. The 5800x is a bit thirstier than the 3700X.

Not sure on options in bios but try running the 5800x in eco mode to suck less power and see if that helps, or an undervolt. struggle to think its the CPU if its only games

varkanoid · 23 Apr 2021 at 12:02

StephenJ said:
If you still have the 3700x you could swap it back to make sure the issues are resolved and that it is 100% due to the new chip

This in the first instance.

welshrat said:
What's your power supply like ? as if its gaming that's causing it could it be lack of voltage somewhere when GPU is drawing more power. The 5800x is a bit thirstier than the 3700X.

Not sure on options in bios but try running the 5800x in eco mode to suck less power and see if that helps, or an undervolt. struggle to think its the CPU if its only games

850W is more than enough for a 5800x. Providing its a decent branded make of course.

I would try it with just one stick of RAM and make sure the Bios is showing the correct timings for the RAM, eg mine defaults to 2133mhz unless I set it to 3000mhz manually. (timings are correct though)

then fire up CPU-Z in Windows and make sure its using the correct mhz/timings

five8five · 23 Apr 2021 at 12:27

Does it blue screen non VR games?

If the other troubleshooting tips don't solve it, do you have another SSD you could install a fresh copy of Windows and test that?

Joxeon · 23 Apr 2021 at 12:34

Have you tried the CMD SFC /scannow just to make sure any system files are not corrupted?

jim_britton · 23 Apr 2021 at 12:37

lots of stuff to try, thanks chaps, wish I had more time to play with this.

I only play VR games but will test non-VR stuff. I agree with welshrat's point that not crashing in Prime95 but crashing in games does suggest a power issue. The PSU is an XFX 850 unit, been solid, maybe it's not any more.

One more data point from this morning. There's a more recent beta bios for the mobo which I flashed. Not crashed yet with this version yet *but* the boost clocks are way lower (and Ryzen Master doesn't appear to be able to read the temperature and package power / current data). With this beta the single core boost speed is about 4600, multi core about 4200 (down from 4800 and 4600 with the older bioses). Hard to interpret that - to me it feels like there's something wrong with this bios that's preventing the chip from boosting properly, but when it does boost it's not stable.

Gah, PCs

. Should have stuck with the 3700x

varkanoid · 23 Apr 2021 at 12:40

Try HWInfo rather than Ryzen Master.
Use it in conjunction with CPU-Z

five8five · 23 Apr 2021 at 12:52

jim_britton said:
There's a more recent beta bios for the mobo which I flashed

I'd recommend you avoid a beta BIOS unless it lists a specific fix that's relevant.

DoneADougalOnSofa · 23 Apr 2021 at 12:55

I had the odd crash when I got my 5800X.
Load-line calibration (LLC) at level 3 on both CPU and SoC seemed to do the trick. Also set RAM voltage to 1.375 from 1.35. (With XMP, of course).

Worth a pop, it'll do no harm if it doesn't work.

philo-sofa · 23 Apr 2021 at 13:14

jim_britton said:
I've had Prime95 running all morning without errors, with and without AVX enabled in 'blend' mode and 'small fft' mode, no errors or crashes

Okay so you've run the memory at JDEC too....

One more idea: the crashing you've experienced does sound a bit like it's failing when returning to a more idle state which is a common issue for Zen 3. Try PBO set to manual and set an all core curve of +15. Out of the box the internal curve the chip is using may be too aggressively negative for the silicon.

welshrat · 23 Apr 2021 at 13:35

Totally agree the power supply should be enough but there are enough tales of people having to have a 850 watt to power the 3080 and maybe it's just on the threshold. Worth a shot to try it. I know I had issues with a Vega 56 and a bequiet 650 a few years back. All fine most of the time but crash under load in games.

jim_britton · 23 Apr 2021 at 13:47

to answer a few of these reponses (thanks for taking the time guys)

Wouldn't normally use a beta bios but this one seems to offer some stability albeit at the cost of some performance

SFC /scannow didn't find any errors

Already using HWInfo, just found it odd that the beta bios prevents RyzenMaster reading some of the chip's telemetry, clearly a bug in the bios

There's no LLC option in this mobo's bios, i'd have to set voltages somewhere else. Don't mind tweaking voltages or curve optimiser settings to extract performance but having to tweak to get stability suggests a faulty CPU or some other hardware issue

Regarding the PSU, replacing this would mean chucking money at the problem. I know there's a chance it's the weak link but it's never given any issues before and it's well within spec for a 3080 + 5800x. I really really don't want to get into a loop of replacing other parts in the hope they fix this. Is there anything that can log the voltages coming out of the PSU in heavy-load situations to establish it as a possible culprit?

bulldog147 · 23 Apr 2021 at 14:06

jim_britton said:
Ayup,

I'm after a bit of advice on a recent Ryzen 5800x purchase from Overclockers.

This replaced an existing 3700x in an MSI Pro-a X570 board with 16 gig of cas18 3600 ram (with an AMD-specific XMP profile), an 850watt PSU, Noctua DH15 cooler and RTX 3080 and various other bits and bobs including a Valve Index.

The system's been solid and stable for ages with the 3700x. Since putting the 5800x in I just can't get the thing to be stable when gaming. It's rock solid doing regular Windows stuff, it always boots and runs brilliantly. But in *some* games (specifically, Automobilista 2, Project Cars 2 and Assetto Corsa Competizione) it'll blue-screen after anything from 5 minutes to an hour. The error is logged in the event log as a "bugcheck 0x00000124" which I believe is a generic hardware fault error code. There are no other errors in the event log (no on-going WHEA errors or anything, just this BSOD).

I can run CPU benchmarks like Prime95 without errors for ages. The CPU will boost to around 4600 all core and 4820 single core when benching, with temps topping out at about 80 when running all core prime85 small tests. Temps in gaming are much lower (in the low 60s). But as I'm a VR player I never actually get to see the temps at the point when it BSODs because I have the goggles on. I also ran the Windows memory diagnostic (no errors).

I've been through some fairly standard trouble shooting steps - set the ram back to default, updated drivers (chipset & GPU), tried 3 different BIOSes (the latest 2 beta bioses and the last 'official' bios). I've also tried with PBO disabled in BIOS and lower TDC, EDC and PPT limits (80, 120 and 120 respectively) but nothing seems to make any difference - it'll run for a while in these games but eventually BSOD with this same bugcheck error.

Is there anything I've missed here? I'm not really sure what to do - should I start the RMA process? It feels like it may be a faulty chip, given that the 3700x has been running like a champ with not a single crash or hiccup for more than a year, but I appreciate that demonstrating that it's definitely the chip will be a challenge

Try disabled C-States and enabled DOCP and leave all on auto and try this !