Processors

[G1n]

Hello, I managed to fix this issue in the past with custom bios settings but after a few months issue is back here. I need help in order to finally fix this or start a RMA.

System spec:

Mobo: Asus Prime 570x-P BIOS 4408
Video: Nvidia GTX 1660 Super
RAM: G.Skill ripjaws 2x8GB 3600mhz F4-3600C16-8GVK bdie
CPU: Ryzen 5800X
HDD: W.D black SSD 512
Windows: Win 10 22H2 OS build 19045

The error:

A fatal hardware error has occurred.

Reported by component: Processor Core
Error Source: Machine Check Exception
Error Type: Bus/Interconnect Error
Processor APIC ID: 10/15/16

I can run prime and all other stress tools, included kombustor for "graphics stress" and system is stable.
Issue happens only while playing certain games, fortnite being the most annyoing making whea error BSDOS in less than 20 minutes. I have never overclocked this machine, it's running all at stock and issue is there anyways.

[cpurpe91]

To narrow down what could cause the issue, do you have another motherboard to test the CPU in?

[G1n]

At this very moment I don't have another mobo =(

[misterj]

G1n, PSU Wattage? Thanks, John.

[G1n]

It is a Super Flower Legion Pro GX 650W power supply, it ran my brother rig for a while with no issues, it had a ryzen 5600X on it.

[misterj]

G1n, I would suggest at least 750 Watt. Enjoy, John.

[G1n]

I think that's not correct. I mean, my brother rig is using a 550W EVGA and he upgraded to a 5800x a few weeks ago, and it doesn't have any whea BSOD. My rig doesn't pull more than 450W even on utopical scenarios, so try to avoid this tricky comments making people buy higher Watt PSU for "spend more money" policy.

[misterj]

Your choice. Enjoy, John.

[G1n]

Need more testing, but I think I found a fix.

[cpurpe91]

Good to hear. Let us know.

[G1n · Accepted solution]

I found some old photos of my bios in my phone "saved them so I could copy settings everytime I would lost them" and realized that the BIOS in those photos was quite old "2802", and I said to myself, wait... I don't remember having any issues back in that time. So the old BIOS was 2802, I tried to downgrade from 4408 which is the latest one, and wasn't able to do so. Contacted Asus tech support they told me to try reverting one by one. I did that but only got as far as to 3405, 3001 and 2802 would throw the "invalid bios file" seems Asus locked how much you can downgrade. I tested the games that generate BOSD with WHEA ID 18, and at least for all the testing I did today, it didn't happen. Also my temps are a bit lower than usual, and there is actually almost 0 stutter compared to the latest one. I even set XMP back on, and didn't got any WHEA, in latest bios having XMP on would lead to BSOD in less than 15 min in some games.
After more testing WHEA ERRORS are gone after downgrading BIOS, I am now able to use XMP and even st tRC to 52 from 85 and system is stable. BIOS 3603 did the trick.

[cpurpe91]

Good stuff my friend. Looks like you just had a bad BIOS. I always say not to update BIOS unless you are having problems.

[G1n]

Yes, lesson learned. I'm not upgrading BIOS again if there is no real need to. Thank you!

[chris-chevyz24]

Mine was doing this in windows 10, I had to turn down the overclock on my ram.

[rumple]

Limit TDP to 140.  

SOC 1.26, IOD 1.1.   CCD 1.01.  VDDP 1.05. Increase CLDO, Chipset and CPU controller voltages by 5%.

Set CPU LLC to 1.

In the memory timings, increase CKE by up to 4. increase TrrdL by 1.  TrrdS by 1. TwtrS by 1. Increase TwtrL by 1.  In subtimings, increase TwrwrScL by 1. Probably the other subtimings as well.  its generally these things that mess you up. you really do need to manually adjust the subtimings except Trdwr, TwrwrSc, TrdrdSc, or Twrrd. (dont touch those)

Make sure BGS swap is off and BGS Alt is on.

Double check CADbus settings , PROCODT, etc, with Ryzen Calc, etc.

Set your Curve optimizer lower, like, oh, 10 on the fast cores and 20 on the rest for ccd1, and 20 on the fast cores on ccd2 and 30 for the rest on ccd2.

Follow my instructions and you will be stable.  miss one step and maybe not.

[Kennbra75]

HP Omen Ryzen 7 5800x

The PC randomly crashes, but not under stress...

A fatal hardware error has occurred.

Reported by component: Processor Core
Error Source: Machine Check Exception
Error Type: Bus/Interconnect Error
Processor APIC ID: 0

Log Name: System
Source: Microsoft-Windows-WHEA-Logger
Date: 2/7/2024 10:43:18 PM
Event ID: 18
Task Category: None
Level: Error
Keywords:
User: LOCAL SERVICE
Computer: DESKTOP-LJBRI66

I don't understand overclocking, so I don't know how to change things.