PC Processors

[CrispyCrunch]

Mainboard: MSI x570 Unify
Mainboard-BIOS: 7C35vA82 (Beta version)
CPU: Ryzen 5900x
RAM: Crucial Ballistix BL2K32G36C16U4B 3600 MHz, 64GB (32GB x2)
Drive: M.2 Samsung 970 Evo+ 1TB SSD
Graphics: SAPPHIRE Nitro+ Radeon RX 5700 XT
PSU: be quiet straight power 11 750w Platinum
OS: Win 10 Pro (64bit) - all updates installed
Chipset driver: 2.9.28.509 (released 2020-11-09)

I first assembled the PC with a Ryzen 3800x a week ago because it was unclear if and when I would get the Ryzen 5900x I ordered. Worked with the included AMD Prism Wrath CPU cooler for one week without any problems.

- Today I installed a Ryzen 5900x and a Scythe Fuma 2 CPU cooler.
- After 20 min the first crash/restart with the following entries in the Event Viewer: WHEA-Logger ID 18 and critical error Kernel-Power ID 41.
- Happens irregularly again and again, sometimes after minutes, sometimes longer: Windows freezes for a few seconds and then the PC reboots. Doesn't matter if load or not.
- CPU temperature between 30 and 40 °C
- Updated to BIOS and chipset driver mentioned above: Problem still exists
- XMP Profile disabled (RAM on 2600 MHz): problem still exists
- CMOS Reset: Problem still exists

Either there is a compatibility problem of something with the CPU, or the CPU is defective?
What to do? Really frustrating.

[Jazzinplayer · Accepted solution]

Im having a similar issue, x570 aorus and 5600x. Have same errors on windows. 

Disable CBP and PBO and run it at default settings (3.7 ghz and xmp on). That works for me. 

[ivan-filippov]

"and for my case helped: SET IN BIOS:
- Global C-state control - disabled
- Power Supply Idle control - set to Typical current voltages."

Yep - I tried the same parameters earlier - this did not help in my case. Unfortunately, the issue with black screen restarts is very transient and unpredictable.  Try to use the pc in 24*7 mode - the longest period without restarts in my case was ~4days - but after that period I got this error. Usually, it can be reproduced much faster ~ 2 times per night.

[artur_aragao]

Esteemed,

I'm still testing.
I acquired new memories and now everything seems more stable. I took one that was inside the mainboard's QLV. ADATA XPG - AX4U360038G17-ST60.

I'm just having problems right now with the AMD VGA driver. Before I tried to play and pow! Reboot.

It is known that there were problems with this driver since version 20.10.2 that have not yet solved this one.

---

@ivan-filippov wrote:

"and for my case helped: SET IN BIOS:
- Global C-state control - disabled
- Power Supply Idle control - set to Typical current voltages."

Yep - I tried the same parameters earlier - this did not help in my case. Unfortunately, the issue with black screen restarts is very transient and unpredictable.  Try to use the pc in 24*7 mode - the longest period without restarts in my case was ~4days - but after that period I got this error. Usually, it can be reproduced much faster ~ 2 times per night.

---

[ivan-filippov]

"I'm just having problems right now with the AMD VGA driver. Before I tried to play and pow! Reboot. It is known that there were problems with this driver since version 20.10.2 that have not yet solved this one."

Are you sure that AMD VGA driver has any relations to these restarts? As I know AMD 5900x has no integrated graphics

I attached system info from my PC with 2 crashes (black screen restart with cpb and pbo disabled and bsod with cpb and pbo on default settings) I do not see amdvga driver in system drivers on my pc

link to these logs:
https://yadi.sk/d/rHEoyg_fRuq2BA

[willeywilson]

Update on mine. Sent the old 5950x back and received a new one yesterday and installed (BG 2049 SUS).  Running Bios Beta 3201 at stock settings was solid as a rock.  Turning XMP and PBO back on didn't cause any issues either. 

I've now managed to get Performance enhancer to Level 2, and Curve optimiser to -10 on all cores (previous CPU required +2 to +6 to work) and it's still stable, zero reboots overnight or when provoking it using the cinebench r23 tools.

To put it into perspective, the old CPU would hit 5050Mhz on 2 cores on CCX0 and wouldn't really go above 4800 on any cores on CCX1.

The new CPU hits 5050Mhz on on 7/8 cores on CCX0 with the last core hitting 5025Mhz and CCX1 hits over 4925Mhz on every single core.  To say I'm happy with the replacement is an understatement.  Oh and with PBO and the negative curve offset I've getting around 4525Mhz all core boosting.

My advice to anyone still struggling is to send it back.  Whilst you can make a dodgey CPU just about stable with positive offsets, why would you want a less than satisfactory product in the first place?

[mackbolan777]

@artur_aragao 

Try using the 20.8.3 driver. I have an entire "fix" based on that driver posted on here but I can't find it right now. Here's the link:

AMD Radeon™ RX 5700 XT Previous Drivers | AMD Scroll to the Windows 10 version you need. Anything beyond the 20.9.2 I can't say will work right. These newer drivers have no new additions for the 5000 series cards as they are EOL. They have things for the 6000 series but for whatever reason the 5000 series components don't behave well using drivers newer than the 20.9.2 and some cards don't like that one, so 20.8.3 is the best bet.

Use DDU in safe mode to remove the current drivers, if you didn't already know that.

[artur_aragao]

Thanks!
I already knew.
I will apply the DDU and validate it.
Much obliged!

---

@mackbolan777 wrote:

@artur_aragao 

Try using the 20.8.3 driver. I have an entire "fix" based on that driver posted on here but I can't find it right now. Here's the link:

AMD Radeon™ RX 5700 XT Previous Drivers | AMD Scroll to the Windows 10 version you need. Anything beyond the 20.9.2 I can't say will work right. These newer drivers have no new additions for the 5000 series cards as they are EOL. They have things for the 6000 series but for whatever reason the 5000 series components don't behave well using drivers newer than the 20.9.2 and some cards don't like that one, so 20.8.3 is the best bet.

Use DDU in safe mode to remove the current drivers, if you didn't already know that.

---

[artur_aragao]

---

@mackbolan777 wrote:

@artur_aragao 

Try using the 20.8.3 driver. I have an entire "fix" based on that driver posted on here but I can't find it right now. Here's the link:

AMD Radeon™ RX 5700 XT Previous Drivers | AMD Scroll to the Windows 10 version you need. Anything beyond the 20.9.2 I can't say will work right. These newer drivers have no new additions for the 5000 series cards as they are EOL. They have things for the 6000 series but for whatever reason the 5000 series components don't behave well using drivers newer than the 20.9.2 and some cards don't like that one, so 20.8.3 is the best bet.

Use DDU in safe mode to remove the current drivers, if you didn't already know that.

---

I could then use this version 2.9.1 which is WHQL and includes these fixes, right?

[majkiel69]

I had rather blue screens with WHEA, not just restarts.

[artur_aragao]

In your case, it may be something beyond memory. You need to investigate. Blue screen is something related to the driver or device. In some specific cases an application that may have replaced an important dll for the system.

Give Malwarebytes a visit beforehand to check if it is clean.

---

@majkiel69 wrote:

I had rather blue screens with WHEA, not just restarts.

---

[bnom]

I currently have this exact problem except with a Asus B550 - A board.

Wondering if I should tweak some voltages or revert Bios back to V2 PI 1.1.9.0

Any ideas? Currently running stock BIOS

[mackbolan777]

@artur_aragao , @bnom 

Yes you can use 20.9.1 if it's working. Even 20.9.2 should work. I found 20.8.3 to be the most stable. Each case may be different and don't but all your faith in WHQL certification. That happens after so much time goes by, regardless of outstanding issues, if nothing is harming the OS per se. Think of it as "period" in a sentence. 

bnom, I need more info than just part of the board model. If you were stable with a previous BIOS, than yes, revert. If you're not overclocking, the 5900x should work just fine out of the box or something else is going on. The BIOS you're listing is 1 version newer than the first stable release. The V2 PI 1.1.9.0 is marked as "Beta" and adds SAM functionality.

New UEFI BIOS updates for ASUS AMD motherboards W01 - B550, X570, B450, X470 - 36 Boards - Includes ...

[bnom]

I currently own the following PC build. It was built about two weeks ago and was running fine but then last week it started having random shutdowns. When the shutdowns start I was running BIOS 1202, I had not updated to a new bios when the shut downs started. I upgraded to 1801 and that did not seem to help.

I have tried everything I could possibly come up with and done all of the recommendations people have given me. I replaced all of the hardware minus the CPU and the issue still existed. I recently updated the chipset and the graphics card drivers and the crashes started happening less on the boot up, but I am still unable to boot any games without instant crashes. 

I am beginning to think my 5600x is faulty but I am unsure how to test this since in OCCT the chip runs fine. I attempted turning CPB and PBO as well as tweaking the CPU voltages and nothing seems to work.

I reset my BIOS to optimized settings and of course that did not help.

CPU: AMD Ryzen 5 5600X 3.7 GHz 6-Core Processor

CPU Cooler: NZXT Kraken X73 73.11 CFM Liquid CPU Cooler ($159.99 @ Newegg)

Motherboard: Asus ROG STRIX B550-A GAMING ATX AM4 Motherboard ($180.00)

Memory: G.Skill Trident Z Neo 16 GB (2 x 8 GB) DDR4-3600 CL16 Memory ($129.99 @ Newegg)

Storage: Samsung 970 Evo 1 TB M.2-2280 NVME Solid State Drive ($139.99 @ Amazon)

Video Card: Asus GeForce RTX 3070 8 GB STRIX GAMING Video Card

Case: Lian Li PC-O11 Dynamic ATX Full Tower Case ($159.99 @ Adorama)

Power Supply: Corsair RMx White (2018) 750 W 80+ Gold Certified Fully Modular ATX Power Supply ($213.99 @ Amazon)

Case Fan: Lian Li UNI SL120 58.54 CFM 120 mm Fans 3-Pack ($80.00)

Case Fan: Lian Li UNI SL120 58.54 CFM 120 mm Fans 3-Pack ($80.00)

[delevero]

I have similar issue with crashing of my 5900x cpu and 6900xt gpu.
I had to revert to the first avaliable bios that made the 5000 series ryzen compatible wiht my motherboard and then i can in 95% of the time play games without it crashing.. actual i played without issues for about a week but it crashed 5 minutes ago in hitman 3.  Any of the new bios updates from december and january for my motherboard = crash within 30 minutes in all games.

So something is clearly wrong with the bios drivers and the new 5900x or 6900xt hardware.

ps.  there are some news about something named AGESA 1.1.9.0 firmware that have af fix of some kind, but its not yet avaliable for my motherboard ( inside the bios )   https://www.techpowerup.com/276833/amd-announces-agesa-1-1-9-0-firmware-updates-improve-fclk-oc-stab...

[zukmeister]

what a sh*tshow this is.  Brand new 5900x high end processor shouldn't have these issues

Any asus guys on here try the beta version 3402 of the bios?  Says it includes an Update AMD AM4 AGESA V2 PI 1.2.0.0

[ivan-filippov]

Hello, guys - a bit more information from me - I send to the shop's service for checking
1. Motherboard (Gigabyte Aorus B-550 Master)

2. Memory (Kingston HyperX Predator [HX436C17PB3K2/32])

3. CPU AMD 5900x

After checks, I got an answer from service - both memory and motherboard had defects - but the CPU worked perfectly fine under load and in idle mode. So if you have the same issue with restarts - try to double-check if the motherboard and memory are fine. Ideally, try to get all parts in one shop with good service - so you can send everything together for quality control in case of any issues.

[Anthos]

Guys try to provide some more information. Just saying you had a crash is extremely vague as literally anything can cause one. Go into your event viewer and see what critical error it states. It helps narrow down a bit what might have caused the issue.

[mackbolan777]

@bnom 

Test your RAM with this: MemTest Manual (hcidesign.com) .  It works better than Memtest86 and is cheaper if you buy the Deluxe version, like I did for $14, Memtest86 is $44. You'll need it more than one time in your life of owning a PC if you do your own troubleshooting. The free versions of any software RAM tester are very limited and most of the time miss intermittent issues. One error means a bad stick, period. Yes, you will pass OCCT and some other things like even Prime95, as mine did. Alternatively you can pull one stick out and see if it crashes, then swap the other stick in, also swap slots (2&4), but it won't tell you if both might be bad.

So in the end of a creeping problem like you're describing, I found that the 3733 RAM and running my IF at 1866 was causing intermittent RAM failures. I needed to use DRAM Calc to find custom timings for running the RAM at 3600/1800 IF and it cured the problems. The problem did happen once more in November and it was a Windows Update that they fixed within 24 hours because it hasn't happened since. Obviously, simply buying slower RAM is easier for those less PC savvy and not all RAM is that tunable, however, money was the barrier at the time. Your issue seems like a bad stick and that speed is fine.

I've had Gskill fail on me a few times and no longer use that brand. Crucial Ballistics, Team Group Dark or Pro, Corsair Vengeance are better choices of brand in my years of experience. That's if you find a RAM stick gone bad.

I would really suspect the RAM, then possibly the PSU with your setup before the CPU since nothing you have done affects the outcome. A final thing is if you mounted the radiator on the front of your case with the hoses up, your CPU could be overheating from the pump sucking air from the top tank. Solution is to remount the radiator to the top, it will take a bunch of rocking motions to get the air bubble out of the tank, it's a frustrating but solvable ordeal I ran into. You can flip the radiator upside down in a front/rear/side mount but most hoses aren't long enough for that. You have a 360mm cooler, I would think it's top mounted, but your case has a few options and if the hoses are up vs. down, check for CPU overheating.

[Funnyjelo]

I have two systems, both ryzen, both under 1 year old. 

System 1 (Main) - Purchased Dec 2020
Ryzen 5900x
ASUS ROG Crosshair VIII Dark Hero Motherboard
G.Skill Trident Z Neo DDR4 3600MHz Memory F4-3600C

System 2 - Purchased Dec 2019
Ryzen 3700x
Gigabyte X570 AORUS PRO WIFI AM4 ATX Motherboard
G.Skill Trident Z Neo RGB 32GB (2x 16GB) DDR4 3200MHz Memory F4-3200C16D-32GTZN

System 1 crashes always, with both sets of DDR4 ram with XMP on.

System 1 doesnt crash with F4-3200C16D-32GTZN and XMP off. 

System 2 doesnt crash ever with XMP on or off. 

Issue on my end is the ASUS Dark Hero or 5900x.

Next test I will do is 5900x swap onto Gigabyte and 3700x onto Asus, identify if it is Mobo or CPU. 

[mackbolan777]

Test your RAM. MemTest Manual (hcidesign.com)

That's the tester I use. Not an affiliate but buy the Deluxe for $14 so you can run it off a CD/DVD/USB. GSkill has had many a failure over my years of building PC's to the point I don't use the brand anymore. Then feel free to swap CPU's because both those boards are decent.

[bnom]

Thanks for this.. I may try to run MemTest Manual just to double check but I think I found the issue.

I decided to change my CPU voltage from auto to offset and offset it by -0.2V. This seemed to stop the crashes, but I am just concerned if I should RMA the chip or if this is a BIOS issue. I was also wondering if I plug in the 6 and 4 pin to the CPU Power headers if this would help. I tried a smaller offset but the crashes were happening at -0.1V

[amas35]

I received my replacement 5900x from Newegg today. It took them two weeks to process the RMA and ship out a new one. If I am reading this lot number correctly, the new one they shipped was manufactured 6 weeks earlier than the one I sent back. I got a new tuf x570 pro board in the meanwhile which still had errors with the old CPU. New CPU posts perfectly fine and is also able to run cinebench. I believe the BIOS is the beta ver that they released on Christmas.

Judging by what else is happening in this thread, it really was an extremely bad bin for my old CPU. I wasn't even able to boot to desktop with the previous chip, unlike others who just needed to pump more voltage to it.

[Funnyjelo]

I have two systems, both ryzen, both under 1 year old. System 1 (Main) - Purchased Dec 2020 Ryzen 5900x ASUS ROG Crosshair VIII Dark Hero Motherboard G.Skill Trident Z Neo DDR4 3600MHz Memory F4-3600C System 2 - Purchased Dec 2019 Ryzen 3700x Gigabyte X570 AORUS PRO WIFI AM4 ATX Motherboard G.Skill Trident Z Neo RGB 32GB (2x 16GB) DDR4 3200MHz Memory F4-3200C16D-32GTZN System 1 crashes always, with both sets of DDR4 ram with XMP on. System 1 doesnt crash with F4-3200C16D-32GTZN and XMP off. System 2 doesnt crash ever with XMP on or off. Issue on my end is the ASUS Dark Hero or 5900x. Next test I will do is 5900x swap onto Gigabyte and 3700x onto Asus, identify if it is Mobo or CPU.

[mackbolan777]

I don't care if the system is 2 days old, check the RAM or waste your time.

[BrigitteB]

Update on my situation:

Motherboard : MSI B550M Mortar Wifi

CPU : Ryzen 5 5600x

Cooler: Noctua NH-U12S

RAM : DDR4 Ripjaws-V 2x8GB 3600Mhz F4-3600C16D-16GVK

GPU : Sapphire RX 6800 Nitro+

PSU: Corsair RM750x

NVME: Samsung 980 PRO 1 Tb

PC built on December 24, fresh Windows install, everything worked perfect for 10 days. I had PBO/SAM/XMP profile options on, not a single issue, temps were good.

Then on one night, as I had been playing a non demanding game for 2 hours, PC starts rebooting out of nowhere (nothing special installed/tweaked that day). I tried to restore Windows, PC would keep rebooting in the process. Finally managed to do so. Now it's hard to even install a driver or check a website without getting a reboot. WHEA-Logger ID 18 everytime. Bios is back to default settings. Tried unplugging and replugging everything. Dunno what to do. Should I return CPU?

Troubleshooting: I took the PC to a local shop, they replaced 5600x with 3600x and everything worked fine.

Fix: RMA'd my 5600x. New CPU arrived yesterday, everything works great so far, no more reboots/WHEAs. PBO and XMP profile enabled. All cores over 4.5 GHz in Cinebench R23 test (scored 11470, CPU temps stabilized at 77 °C).

Bios version: 7C94v14

[delevero]

Try to install only 1 ram block in your computer  ( be aware it matters in what slow you install it, so if computer wont boot afterwards move the ram 1 slot and try again.
( i have 32gb but i removed 1 so i only had 16 and that after a short wile of testing resulted in my computer no longer crashing in games.. well at least so far i only played for 2 houres since i removed 1 ram stick but i would normal have a crash but i dident so far )

[JoltCola]

I don't see any particular reason to purchase either HCL memtest or memtest86. Just run the latest free version of memtest86 booting from a USB stick. If it runs for an hour with no errors you're 99% likely to be fine. Run overnight and round up to 100%.

Everybody should do this whenever they buy new RAM or put old RAM into a new motherboard. It's shocking how often RAM can be faulty, or your new motherboard isn't driving it properly, and RAM can cause all sorts of intermittent difficult to purposefully trigger weird instability.

Windows also comes with a built-in memtest, run "mdsched" from an elevated command prompt. It's probably fine too, but a bit of a pain in the butt as you need to boot to windows desktop, run it, then reboot every time. Easier to just boot off a USB stick with the free version of memtest86.

[Juannox]

new ryzen user here after more than 25 years with intel... 
5 days ago I built my new AMD system, and from the first moment I have a lot of whea random errors, and some unwanted reboots with no BSOD... I have tried everything from this forum but nothing works, I tried F31 bios version from the first day, 2 days ago I upgrade to F32 and 1 hour ago I Upgrade to F33a but 30 minutes later, again the same error 

I also did a memory test with memtest86 using a bootable usb (more than 5 hours) and everything went well

Proc Ryzen 5900x, (stock, never OC'd) 
MB: Aorus X570 Elite Wifi (Bios version: F31,F32 and F33a)
Mem:  HyperX Fury 32GB 3600MHz DDR4 Ram CL17 DIMM (Kit of 2x16GB) - I tried XMR 1,2 and off but the problem persist
SSD: Intel Optane 900p 280 GB

whea logs since the last os fresh install: https://imgur.com/b55Jtns

I have no idea what to do....

PD: from everything I have found in multiple forums, most people with this problem have 32GB 3600MHz, It is not a conclusion but it is an interesting fact that catches my attention

[ivan-filippov]

"I have no idea what to do..."

Juannox, the issue can be in PSU\Motherboard\CPU\Memory  - you need to retest all parts in a service that has required spare parts for this check

[JoltCola]

Usually newer AGESA fixes reboots on idle, but I see F33A has AGESA 1.2.0.0. Try setting power supply to typical in CBS settings. Also try giving your vCore, vMem, and vSOC a bit more juice via offsets, see if any of those stabilize it.

[artur_aragao]

I will spend a few things here that I have been analyzing and have greatly improved the stability in the system.

These WHEA problems can be related even with PSU, Mainboard and CPU. I particularly consider that the problems can in most cases here be linked to other factors.

If you have tested all the components, we can predict that there are more problems related to the devices (not the ones already mentioned) connected to this system, or the system itself and drivers that may be contributing to it. Drivers use IRQs transparently for us.

In my case, it stabilized A LOT by changing memories that were not compatible and uninstalling Corsair Link, which is no longer supported and therefore, the app may have problems that could cause BSOD, crashes and restarts. Take a look at the Corsair Forum and you will see that. I haven't fully remedied the problem, but it has been 4 to 5 days since I received BSODs. The Dragon Center from MSI I do not enable the gamers features and so far the system has not caused me any more problems.
However I am analyzing.

I am suspicious, also that it may be a bug in the Microsoft operating system. But that would need to be analyzed. I disabled Fast Boot in Windows 10 and in the BIOS. It helped a lot.

So the tip I give, if you have devices that are constantly monitored and managed by some software, validate that the USB cable that is connected or the app is not causing the anomaly.

And I consider that there is still a fine adjustment of BIOS for these problems.

[bnom]

Just curious, how long did the RMA process take?

[artur_aragao]

Update to ComboAM4PIV2 1.2.0.0 for the MPG B550 GAMING EDGE WIFI is out.

Validate with your manufacturers if the BIOS came out.

[zukmeister]

I think i'm going to start an RMA on my 5900x... my last time buying an AMD processor.  What a shame 

[artur_aragao]

Good!

I had already achieved stability with some of the things I did and posted here. I was still analyzing if anything could happen, but now the BIOS that was waiting is out. The ComboAM4PIV2 1.2.0.0 version seems very stable. However, testing still. I already activated Resize-BAR. We will test later with the games.

Validate with your mainboard manufacturers if it already exists, even if it is a BIOS BETA with AGESA ComboAM4PIV2 1.2.0.0. I felt an improvement here in stability and performance.

I hope you find it soon

[ivan-filippov]

"my last time buying an AMD processor.  What a shame"

zukmeister, I also blamed the processor but after checks in the shop's service, I was told that the defects were in the motherboard (Gigabyte B550 Aorus Master) and in the memory (Kingston PredatorX 2x16GB 3600c17).

I mean, that I understand all your disappointments but let's try to be cold minded here:

1. You will not find any Intel's CPU comparable to 5900x\5950x (and it's useless to wait for Rocket Lake as 11900k will have 8 cores only)

2. It's ok if sometimes we got hardware with defects - that's how we can test the support service

Anyway - I hope that your next build will work perfectly. Good luck

[Sbanks1987]

The symptoms I was getting were very similar to this post. Tried everything and was putting it down to faulty 5950x or motherboard. In the end I've not had a similar crash since I uninstalled the nzxt cam software. Maybe it's worth trying with similar software like someone mentioned above with the Corsair. Hope it works out because I understand how frustrating it can be. 

[artur_aragao]

I'm glad you solved your case.
I was the one who posted about Corsair Link.
I recommend that you do a search, as most likely it is actually some software installed.
But I need to make a reservation. The BIOS with AGESA 1.2.0.0 gave a super stabilization in the system as a whole. I understand that this can also be a resolution step.

I still have Fast Boot in BIOS and Windows 10 disabled. The Fast Boot in the BIOS kept my SSD speed below the nominal. And these disabled features have further stabilized my situation with the previous BIOS. So, I still need to take a few more days and validate the Fast Boot feature again.

[mackbolan777]

I use corsair Link and could never pin an issue on it relating to Error 18. It doesn't even work until the PC boots into Windows. Mine controls the H110i fans and keeps the pump at 100%, controls the DPI/lighting on my Harpoon mouse and if I have the VOID plugged in, it controls the lighting, equalizer preset, and volume. I have it set to not be read by HWiNFO because of alleged problems. Fast Boot I always have disabled since Windows 10 came out. Learned the hard way that if that's on and you need to get into BIOS or stop an OC at boot, you're "bleeped". It also skips loading certain drivers one might want loaded and skips pre-boot testing that might catch an issue before the OS loads.

[KenAdelaide]

mate same issue here!

Mainboard: ASUS x570plus
Mainboard-BIOS: 3402(beta, latest one)
CPU: Ryzen 5900x
RAM: corsair 3600 MHz, 8GB X 2
Drive: M.2 Samsung 970 Evo+ 1TB SSD
Graphics:  AMD 6800XT
PSU: greatwall 1000W
OS: Win 10 Pro (64bit) - all updates installed

I disabled PBO, but running my ram at 3200MHZ, the problem still exists 

someone on the Internet solved this issue by setting VDDP、VDDG CCD/IOD at 0.9V but it doesn't work for me

others said they solved this issue by contacting AMD to get a new one, they claimed CPU with WHEA ID18 error is a defecated CPU... I'm going to contact local store tomorrow, but I'm not sure if they are willing to give me a new one or not

[jaydeejoy]

exatly same issu with same mobo and same Ram
how  i fixed it ( 20 days stabile no restart)
i sold corsair 2*8 3600
bought crucial 2*16 3200
updated lastest bios
and
UNCHECK fast boot on windows power plan

XMP ON
PBO ON

3900x + tuf x570 + 2070s+2*16 32 Gb 3200

Try it.