PC Processors

[flagrant99]

Watching a video on YouTube in chrome my computer rebooted yesterday and the following was in the event log:

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

Dump shows

BUGCHECK_CODE:  124

BUGCHECK_P1: 0

BUGCHECK_P2: ffffe28909ef18c0

BUGCHECK_P3: faa00000

BUGCHECK_P4: 80b

PROCESS_NAME:  smss.exe

IMAGE_NAME:  AuthenticAMD.sys

This happened again later in the day while I was away from my machine. 

This is a new build. Fresh windows 10 Install. All latest drivers from AMD, etc.

Today while I was in File Explorer in windows 10. I get:

reboot.

A fatal hardware error has occurred.

Reported by component: Processor Core

Error Source: Machine Check Exception

Error Type: Cache Hierarchy Error

Processor APIC ID: 3

BUGCHECK_CODE:  124

BUGCHECK_P1: 0

BUGCHECK_P2: ffffb5826f8ed030

BUGCHECK_P3: bea00000

BUGCHECK_P4: 108

PROCESS_NAME:  smss.exe

IMAGE_NAME:  AuthenticAMD.sys

Specs:

MOBO ASUS ROG Strix B550-I Mii ITX Mobo Firmware 2007 Beta (Lately they are all Beta so do I rollback to a non-beta?)

CPU Ryzen 5 5600X Fabric Clock @1600 DOCP Profile 1.

RAM 2X16GB G Skill Trident Z Royal F4-3200C16-16GTRS 1.35Volts 3200MHz

GPU XFX Radeon RX 570 4GB

SSD Samsung - 980 PRO 1TB PCIe 4.0 NVMe M2

AsusRyo CPU Water Cooler (CPU Temps Hit 55c max in prime 95)

PSU CORSAIR - RMx Series 850W 

Tests Run so far:

MemTest86 V9.0 Free (64-bit) passed twice with RAM @AUTO-2666 and @DOCP-3200.

Running Prime 95 now for over an hour in windows now. so far no problems. 

Conclusions so far:

Problem has NOTHING to do with load. In fact it likes load. More likely to happen when machine is idle doing barely anything. 

[flagrant99]

I checked and Bios 0207 is Update AMD AM4 AGESA V2 PI 1.2.0.1 Patch A. 

Gonna try with DOCP off so ram is at 2,666 and see if I make it the day without a reboot. 

My gut is telling me it's not the ram since memtest is passing at all speeds.

Maybe my Fabric Clock can't do 1600 when idling?

[flagrant99]

Updating Armory Crate it just crashed again. Disabled Precision Boost Overdrive and PBO Fmax Enhancer. Crashed again. 

[flagrant99]

I can crash on demand now. Asus Armoury Crate -> Update Center -> ROG RYUO -> Update. Crashes half way every time. The RYOU is AIO Water Cooler with screen connected to USB 2.0 Header on MOBO to set Screen Display.
Guessing it is trying to communicate over USB to it.

Swapped RAM. Still crashes.
Downgraded to last NON Beta BIOS Version 1803, AMD AM4 AGESA V2 PI 1.2.0.0. Still Crashes.

[flagrant99]

Swapped GPU 

GPU XFX Radeon RX 570 4GB

to 

an old XFX Radeon HD 5450 512MB DDR2

It worked!

Doesn't look like I can run the ROG AIO Cooler update again so I lost my reliable test for failure. 

So does this mean RX 570 not compatible with 5600X? I can't buy a new video card now to try something else.

Guess I can try reseating the RX 570.... 

[flagrant99]

Well the old Radeon HD 5450 maxes out at 1024*768. Pretty much unusable. So I went back to my RX 570 @ 4k. Reseated, Swapped Modular power connector, Used different ports, etc. 

10 minutes in reboots while look in Event Viewer  for older WHEA errors. Made another one. 

I have no idea what to do now. Can't buy a another video card. 

My Radeon RX 570 has been running fine with my intel 3770K for years. 

It's not the video card on intel.

It's the Radeon RX 570 video card on AMD Ryzen 5 5600X. 

[flagrant99]

Upon further reflection the XFX Radeon HD 5450 working could have been a fluke and it's not exactly taxing anything running 19 watts @ 1024*768.  

So I think it has to be CPU or Mobo (PCI Chipset). I pulled my AMD Ryzen 3200g running on my old b450 and am trying it with the RX 570. No reason I have to use the APU. So far no crashes but the night is young. Gonna RMA the CPU or the Mobo depending on whether I get crashes or not. 

[flagrant99]

My 3200g runs SOLID with RX 570 as GPU. NO reboots. Zero.
So it is definitely something wrong with my 5600X CPU.

RMA Service Request: ticketno:[8201076953].

Now I wait.

[flagrant99]

As part of RMA I had to reseat my 5600x to get pic with RMA # written on paper.
I guess it makes sense to try reseating cpu in case some dust or something got in socket.
Since I reseated I retested Thursday, and I was almost convinced it was OK at Stock non D.O.C.P Ram Speed (1067).
Ran 4 hours straight no reboots.

So I felt bold enough to try D.O.C.P. F-Clock from 1067 to a whole 1600! Brought up Youtube. Reboot.

Started thinking I needed to raise SOC voltage to 1.1 to get RAM stable for rated 1600 F-Clock speed.
No'p reboot.
Back to non docp speed with Fabrick clock @1067 Mhz, ram @1067 Mhz.
Reboot.
Youtube.
Reboot.

I have NO interest in an Overclock, but is 1600 Fabrick clock an OC to my ram @ rated speed?
Doesn't matter since it can't handle @1067 Mhz F-Clock reliably anyway.

Whatever is causing this has no pattern, except that no load causes it.
CPU ram tests, Prime 95, etc. max CPU speeed it runs solid.
It's happy with load. It likes the heat.
Stress tests are a total waste of time.

Youtube and Notepad are the new tests for stability on my 5600x.
idle almost nothing going on for all those cores.
That's what causes the reboot.
And it'll work every now and then.
Sometime for hours.

I'm back right now on the 3200g. My 3200g shows no Fabric Clock in Ryzen master, but DOCP up running RAM CLOCK @1600.

YouTube works solid.
2 days no reboots on the 3200g (Using RX 570 GPU)

5600x is going back.
I sent pictures Thursday night.
Waiting.

[flagrant99]

I was going to send my chip back as part of RMA# 2000310985, but I found my packaging, so I returned it to best buy and ordered a 5800x. I have also since purchased a 4700g (to upgrade my 3200g) and that also runs fine in this 550 mobo.  When the 5800x comes in will update the saga.

[weenan]

Remember reading somewhere on this forum about something similar where the culprit turned out to be using an older psu with the Ryzen. He also always got crashes when watching youtube during low load.

With older generation psu it could be fixed with BIOS setting for the Power Supply. Check you BIOS for "Power Supply setting" and set "Powers Supply Idle Control" to "Low Current Idle". Could be worth a try.

[flagrant99]

Just scanned my mobo bios manual for that setting. Don't see any power supply settings. If the 5800x is unstable will look harder. Should be in by end of week. 

[weenan]

On my MSI MAG X570 Tomahawk it´s under OC - CPU Features.

I think I was wrong in my previous post because now I´m unsure if it was to be set to low current idle or normal. But if you find it try changing it and see if that does anything.

[flagrant99 · Accepted solution]

I've had my 5800x a few days and it works fine. No instability. 

[Bica]

Did you just write the story of my last 3 days half a year in advance? 

I've been pulling my hair over this with my 5600x. blamed my psu, gpu, motherboard. 


On stock bios settings the only sure way to keep it from restarting is to have prime 95 running in the background.
Can get it to run stable'ish' if i set vcore to 1.15v but then it will never underclock on idle and i feel i lose a lot of performance at the top too. (It also doesn't get hot at all anymore then. like 60 degrees on load) 

was already kinda dreading it but when i read this i called the place where i bought this and have em take a look at it. If they experience the same issue i can get my money back or a replacement.


My first amd component in years after being an intel fanboy for a decade and a half and it's not going well so far..

[flagrant99]

Yean man send it back. I am convinced something was wrong with my chip. Bestbuy took mine back no questions asked. I feel sorry for anyone that gets that chip. Load tests all will pass.  All I could get at the time was a 5800X to replace it which cost more. At least I got my sanity back. Still Now I fear that pain so much I am afraid to update my bios. Not updating anything since mines been running solid for a long time now. Very happy with my 5800X. 

[majorscrewball]

Hello.  Did lowering the voltage resolve the issue?