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. 

[Genolandia]

My PC is a 5950X and a 3090 ROG STRIX OC. PC currently running well with the 5950X being replaced termporarily with a 3800X. Everything fine and dandy right now. I am willing to bet it is the CPU since the windows install was really unstable with the 5950 and flawless with the 3800. I reinstalled as an experiment and it seems I was correct. With the 5950X I had restarts even in the Windows install that I did not have with the 3800X. Not a single crash since I changed yesterday morning (almost 24 hours).

Perhaps there is interplay, but I doubt the GPU is causing the idle crashes since it is not used much then. Still an interesting and useful observation.

[Genolandia]

This is certainly not solved mods. You should remove that label. Those "solutions" are band aid fixes, not proper solutions.

[colesdav]

Nothing changes then...

Tuesday

Re: Ryzen 5900x: System constantly crashing/restarting WHEA-Logger ID 18 and critical error Kernel-P

It is probably CPU frequency boost and variation causing enough noise to cause DRAM errors and PC Crash. 
I find it better and more stable to run my PCs with a fixed manual CPU overclock and BIOS voltages adjusted.
I run the DRAM at the fastest rated speed possible for the motherboard. 
I  turn all of that AMD "CPU Boost" nonsense off, because of the crashing it caused. 
That is my experience with it all. 

[colesdav]

It is probably CPU frequency boost and variation causing enough noise to cause DRAM errors and PC Crash. 
I find it better and more stable to run my PCs with a fixed manual CPU overclock and BIOS voltages adjusted.
I run the DRAM at the fastest rated speed possible for the motherboard. 
I  turn all of that AMD "CPU Boost" nonsense off, because of the crashing it caused. 
That is my experience with it all. 

[860lacov]

Well

I don't want to overclock st all

For me solution when cpu has some idle clocks that go higher under load it's a perfect solution.

Like in intel turbo boost.

J just wanted to have everything in bios on defaults and I believe by default there is 3.7 base clock and it should go up to 4.8

While buing cpu I was sure that those parameters are stable amd defaults.

[MishkaKosolapy]

Count me in as yet another person who has been hobbled by the kernel-power (event ID 41) errors with a Ryzen Zen 3 / 5000 series processor. 

Here are my current hardware specs:

AMD 5600x / MSI MEG Unify x570 / 32 GB G.skill 4000 Mhz RAM F4-4000C16D-32GVK (currently clocked to 3800 Mhz) / AMD R390  GPU (waiting to upgrade when GPUs are in supply) / NZXT X63 AIO cpu heatsink / PSU NZXT C850 850 Watt / Sabrent 1 TB NVMe PCIE 4.0 SSD

Randomly, my computer will crash with a blank screen, fans will temporarily increase speed, and then I have to reset to get it back up.  Windows logs a critical event 41 each time (kernel-power). 

I actually started off with a Asus Rog Strix x570-F motherboard, and I had the exact same problem.  I switched out the motherboard to the unify thinking the motherboard was the problem. When I was using the Asus, I tried 4 different BIOSes, up through the latest version 3001. None of them had a noticeable impact on the crashing problem.

What I did notice, both with the Asus and MSI boards, is that the problem tended to increase in frequency with higher RAM speeds. The problem was also more likely with a gaming benchmark like 3DMark rather than more pure CPU loads, like geekbench. That led me to believe that RAM was the problem, so I set my RAM at default 2133 Mhz and also tried an old 8 GB kit at 2400 Mhz.  With both types of RAM running at slower speeds, the system was more stable, but I could still induce it to crash by running 3DMark (it would crash 1 out of 3 runs vs nearly every time at high RAM speeds). The crashes were very frequent for me—with normal usage they would take place 2-3 times per hour for high RAM speeds and 2-3 times per day at low RAM speeds.

I have followed the advice on previous posts and turned off Core Performance Boost. Since then, the system has been completely stable, regardless what RAM speed I use. So the problem is clearly the CPU, but it seems to be related to its boosting capability and how it interfaces with the RAM, since higher RAM speeds aggravate it.

Other people have had luck tinkering with voltage, current, and PBO to get their systems to operate stably with CPB enabled. I don’t really want to mess with settings that aren’t covered in the warranty in order to get my processor to work when it should be able to operate stably at default settings. I don’t want to have to live with a dudd. So I’ll initiate an RMA once all of the shipping craziness passes after the holidays, assuming AMD doesn’t completely resolve it through an AGESA update by then.

PS - Since the original poster said in a later post that the RAM voltage increase that he has marked as a solution hasn't fixed his problem, I recommend marking this thread as unresolved.  

[860lacov]

I will add something.

Everything on default CPB auto

Pc reboots in games and blender benchmark

CPB disabled

Pc is stable but it looses Bout 25% if performance (in blender benchmark)

Shi...y thing.

Oh. One more thing.

Everything default bios shows vcore over 1.4V

CPB disabled - vcore around 1V

Without  CPB is a waste of money i think 

[fastjoe90]

Spent a few days testing most of these fixes, except disabling CBO as I feel like that totally defeats the purpose of upgrading CPU. Windows started a cumulative update and my pc rebooted part way through. Had to do a dism restore health. That was the last straw. I returned the 5900x to Microcenter and luckily grabbed a 5800x on the way out and have had zero reboots or WHEA 18 since on the exact same motherboard, bios, xmp, etc. 

[860lacov]

Totally agree but 5800 is not 5900 so not the best solution but many the only one

[ivan-filippov]

Having the same issue AMD 5950x. I got random reboots even with turned off core performance boost and on non-xmp memory

why we do not see any feedback from AMD support here?

[daya_papaya]

Not a good feeling when the only reliable solution to this problem you've seen in a few months is "just buy different CPU/GPU"

[860lacov]

True

Today I ordered 10850k and going to return 5900x

"Sad but true"

[usurper]

if your crashes happen at low load, and no downclocking helps, i'd once again suggest to read what i posted in the other thread

https://community.amd.com/t5/processors/new-ryzen-3900x-x570-random-restarts-whea-logger-error-id17-... 

looks like PSU can shut down if the power load is too low, and MBs have options to not go lower than some minimum value. in addition to "power loading", there's also an "power on idle" CPU-specific setting, which can be set to "typical" instead of "low".

[amas35]

I have had this issue for a couple months now with my 5900x and was blaming my RAM since it was an extremely old 2133 mhz kit. I just received my new kit today and am still having these hard crashes. At the beginning, I was able to at least boot into Windows but crash on any RAM heavy tasks but as of a few weeks ago it would not get past the windows load screen before hard crashing.

I have tried many settings and voltage changes suggested around the net to no avail. The only stable """solution""" I have found for myself was to change the Core Leveling Mode in my Asus bios from auto to Three which basically turned off half of the cpu cores. The computer finally worked doing this and didn't crash under heavy work/gaming/stress loads but it clearly isn't a viable solution.

Some unstable changes I have tried, that was at least able to get me into Windows, was to set a static cpu ratio/voltage. This led to various issues inside Windows such as temperatures randomly climbing 100c+ in stress tests. Another weird one from static voltages was the CPU idling at 1-1.5ghz then proceed to lower itself to .5ghz during any kind of load.

I am at a loss of what else I can do to make this a usable chip aside from disabling half of the cores. It seems like many of you here were having similar issues too but were able to solve through other various methods. Seems like it's going to be an RMA or a replacement from Newegg for me.

• 5900x
• Asus tuf x570-plus (wifi) (bios 3001 Agesa ver combov2_1180)

[amas35]

Tried again with new beta BIOS and still nothing. I'm going to get a replacement from Newegg and a new motherboard while I wait.

[860lacov]

SOme time ago I did quite long memtest. No errors.

Checked xmp on and off.

And I have reboots.

Last hope was to change memory. I returned Hyperx CL17 3600 and changed them to G.Skill 3200 CL16

After change, I was able to complete 3 blender benchmarks in a row. Couldn't do that before. I was able to play Witcher 3 for over 10 minutes. Couldn't do that before.

It's Christmastime, so I didn't want to abandon my family for my PC. Today (precisely tonight) I had little time to check everything again.

Started blender benchmark and PC rebooted again.

Hate AMD right now. I wanted to return CPU but new copy is hard to get and prices are about 240 USD higher so if I will return my CPU i won't be able to buy a new one for the money a can get.

After everything I tried I'm starting to doubt that it is a bios problem.

If there are good copies and bad copies of this CPH then how could bios programmers distinguish one from another?

And I wonder. People here usually have a problem. With idle. I have biggest problem in blender benchmark test 6 (so last, most demanding one)

[Unodediez]

Im having the same issue with a 3900XT i dont know what to do, should i RMA the cpu?

[860lacov]

And this is very bad news.

I returned 5900x with hope that 3900x or xt will work without any problem.

Which bios version?

Maby you could try bios before ageisa path c?

[Unodediez]

I tried with stock bios version, with the last stable (3003) and with the last one (beta version) 3103. Same result on all. 

Btw I don't know what is the agesa path c. I'm completely noob on this

[860lacov]

I don't remember since which version ryzen 5000 are supported.

I'm just guessing that bioses with ryzen 5000 support can have some problems.

I couldn't tell if my cpu or bios was faulty. That's why I returned 5900x.

I just wanted to say that you could check last stable bios without ryzen 5000 support.

what motherboard do you have?

[Unodediez]

I have an Asus Rog crosshair hero VIII wifi

[860lacov]

I would try bios Version 2206 

I checked on asus website and I believe that this was the last version before big update.

I can't promise that this will help but it shouldn't make your situation worse ;)

If this will help please let me know.

As I said. I returned my 5900x because I wasn't able to tell where was the problem in my system.

[jpee80]

I sent AMD Support the reddit post where this thread is included. I hold my expectations low but someone up there needs to say something about this. 

[860lacov]

Probably not only you :)

We are on AMD community forum. It seems to me that AMD knows already and don't want to answer. If this is true then this problem could be more serious.

[Unodediez]

@thunk_stuff everything fine with the new cpu?

[thunk_stuff]

@Unodediez Yes, been stable. No unexpected crashes or WHEA errors in Event Viewer.

Currently running auto overclock, curve optimizer to -20 (CCX0) and -25 (CCX1), +25 mv core offset, +100 Boost. It did crash when I undervolted too much, so that's why I bumped +25 mV core offset. Been running over a day now at these settings. If it crashes again I will tweak voltage until I get something long term stable. I read you want to test for a least a week or two.

Getting 700 (single) and 10035 (multi) in CPU-Z.

Cinebench R20 is 638 (Single), and 8631 (Multi).

When running Cinebench R20 Multi Core, average core voltage is 1.205 and temperature is 64c (Noctua NH-U12A).

[tim716]

---

@thunk_stuff wrote:

@Unodediez Yes, been stable. No unexpected crashes or WHEA errors in Event Viewer.

Currently running auto overclock, curve optimizer to -20 (CCX0) and -25 (CCX1), +25 mv core offset, +100 Boost. It did crash when I undervolted too much, so that's why I bumped +25 mV core offset. Been running over a day now at these settings. If it crashes again I will tweak voltage until I get something long term stable. I read you want to test for a least a week or two.

Getting 700 (single) and 10035 (multi) in CPU-Z.

Cinebench R20 is 638 (Single), and 8631 (Multi).

When running Cinebench R20 Multi Core, average core voltage is 1.205 and temperature is 64c (Noctua NH-U12A).

---

Dude, your settings look like an abomination. 

You decrease the core voltages with Curve Optimizer and increase them with the core offset.  The horses in your harness pull the opposite ways. :) 

Just remove the core offset and decrease the negative Curve until the CPU works stable.

[Erzkonservativ]

Hi, made a new account since I seem to have solved my problems and wanted to share what I did.

My system:
5900X, B550 Tomahawk, RTX 3080 FE, 2x8GB from G.Skill (they can only do 3200Mhz, so shouldn't be a problem)

Issues I had:

- Kernal 41 errors (PC would reboot out of nowhere on stock BIOS settings)
- with C-states enabled this would happen within 10 minutes
- disabling C-states stretched that time to 30-60 minutes
- happend in idle and some games (never in Cyberpunk though)
- setting a EDC limit of 200A and the curve optimizer to +8 made the system stable, but that seemed like a bandaid solution. Still big thanks, I used those settings from a post from here!

My system now seems to run (ran for about 13 hours yesterday without issues, but of course I could've just been lucky) stable on 100% stock BIOS settings. Here's what I did:

- Uninstalled old Intel drivers with DDU. Read somewhere that this can help and my Windows installation actually ran an Intel system years ago. My 2600X never had any problems with that, but maybe the 5900X has?
- Flashed my BIOS yesterday with the new 7C91vA53 version (that has ComboAM4PIV2 1.1.9.0)

I did not try if the DDU uninstall of the Intel drivers already solved the problem. Still ran the system with the bandaid settings for about a week after that. So I am not sure which of those steps solved my issues. Maybe both. Will update should my stability change again.

[JoltCola]

I betcha it was the 1.1.9.0 AGESA. It is a huuuuuge improvement on stability, particularly on CPUs with multiple CCDs.

[Erzkonservativ]

Yeah, I would think so too. Still wanted to include the DDU thing since there is a small chance that it had an impact and it's very easy to do.

[860lacov]

I returned my 5900x few days ago and just few hours ago I ordered one again. It was in stock for hours or something.

I had to pay more but now I'll be able to tell if it was mobo/ bios or cpu.

I have msi tomahawk x570 so no new bios yet.

[jaydeejoy]

---

@Erzkonservativ wrote:

Hi, made a new account since I seem to have solved my problems and wanted to share what I did.

My system:
5900X, B550 Tomahawk, RTX 3080 FE, 2x8GB from G.Skill (they can only do 3200Mhz, so shouldn't be a problem)

Issues I had:

- Kernal 41 errors (PC would reboot out of nowhere on stock BIOS settings)
- with C-states enabled this would happen within 10 minutes
- disabling C-states stretched that time to 30-60 minutes
- happend in idle and some games (never in Cyberpunk though)
- setting a EDC limit of 200A and the curve optimizer to +8 made the system stable, but that seemed like a bandaid solution. Still big thanks, I used those settings from a post from here!

My system now seems to run (ran for about 13 hours yesterday without issues, but of course I could've just been lucky) stable on 100% stock BIOS settings. Here's what I did:

- Uninstalled old Intel drivers with DDU. Read somewhere that this can help and my Windows installation actually ran an Intel system years ago. My 2600X never had any problems with that, but maybe the 5900X has?
- Flashed my BIOS yesterday with the new 7C91vA53 version (that has ComboAM4PIV2 1.1.9.0)

I did not try if the DDU uninstall of the Intel drivers already solved the problem. Still ran the system with the bandaid settings for about a week after that. So I am not sure which of those steps solved my issues. Maybe both. Will update should my stability change again.

---

whats your RAM part number?
can they work XMP/DOCP enable mode
are they 3600 kit?

[Erzkonservativ]

The RAM part number is F4-3200C14D-16GFX

So only 3200. 

I don't use XMP and used "custom" settings with Ryzen DRAM Calculator instead.

[jaydeejoy]

---

@Erzkonservativ wrote:

The RAM part number is F4-3200C14D-16GFX

So only 3200. 

I don't use XMP and used "custom" settings with Ryzen DRAM Calculator instead.

---

ok ty.
did u try 3600 ram kit?

[jpee80]

I finally got a replacement and the issue is fixed.

The new CPU's BG#: BG 2045PGS

[Anthos]

I'll try to give a quick update about my case, I' ll start from the beginning again to make it easier to follow:

5950x, Dark hero, 3600C16 ram
-put the system together 2 weeks ago, no fresh windows installed, just put everything together and carried on using my previous intel/ddr2 windows
-first week everything solid, no single issue
-installed windows last Sunday on a secondary drive just for benchmarking purposes, multiple WHEA errors, even got one (I assume) when booting off the usb bios to install the windows
-bleep up on my behalf accidentally installed on wrong drive and lost my old windows installation
-within the next 24h on the fresh install during random idle etc system would reboot, at times back to back, while others spaced out (I was still on stock bios)
-update to latest stable bios was still the same, load default values, no change, still random reboots
-disabled c-states and updated to the latest beta bios, seemed stable
-no WHEA for the next 24h
-decided to re-enable c-states, ram on DOCP, and re-enabled CPO, CBP
-still no WHEA error in the next 24h
-escalated further by messing around with more DRAM overclocks (beyond DOCP), all core overclocks etc etc
-still no WHEA errors

So pretty much I had a fully stable system for a week, for whatever reason went ape**bleep** for 24h and threw all kinds of WHEA errors at me and then re-stabilized again... go figure...

[JoltCola]

Yep, AGESA 1.1.9.0, in most manufacturers' current beta BIOS releases, fixed WHEA on idle/near-idle for me and many others. Normally I would run away from a beta BIOS like it's on fire, but in this case I really do recommend it.

[Anthos]

---

@JoltCola wrote:

Yep, AGESA 1.1.9.0, in most manufacturers' current beta BIOS releases, fixed WHEA on idle/near-idle for me and many others. Normally I would run away from a beta BIOS like it's on fire, but in this case I really do recommend it.

---

Yeah, the thing though is... Why didn't I get any WHEA errors the whole first week with the stock bios? It all seems weird...

[860lacov]

Well

Not everyone got this problem.

If you don't have any errors / reboots then you probably won cpu lottery.

Usually it means that one can OC cpu more.

This time AMD f...d up

Usually

Good cpu = good oc

Bad cpu = bad oc

This time

Good cpu = stable system

Bad cpu = unstable system and errors

[minos8]

I have same problem, 5900x with MSI B450 PRO CARBON AC last beta BIOS.  

 swapped only the processor from a 3700x all my setup is good. AMD chipset drivers updated.

BIOS
Version 7B85v1F1(Beta version)
Release Date
2020-12-22

Problem
Black screen crash if I do Cinebench r20 multiples times.
Black screen crash after some minutes when gaming.

Tried all settings found on internet, the only one who works BUT IS NOT A SOLUTION
Disable CBP and PBO and run it at default settings (3.7 ghz and xmp on).

What is next ? I opened AMD RMA , wrote on MSI Forum, and I am already tired of this.

[BrigitteB]

Motherboard : MSI B550M Mortar Wifi

CPU : Ryzen 5 5600x

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

GPU : Sapphire RX 6800 Nitro+

PSU: Corsair RMX750

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?