PC Processors

[rlauzon]

Hello,

I am having issues with two AMD system fresh builds, bot are receiving BSOD error "clock_watchdog_timeout error"

System 01:

Motherboard: Asus PRIME B450M-A

BIOS: 2.17.1246

Processor: AMD Ryzen 5 3600X 6-Core

Memory: Corsair Vengeance LPX 32GB (2x16GB) - CMK32GX4M2B3200C16

Video Card: EVGA GeForce GTX 1050 Ti GAMING, 4GB GDDR5, DX12 OSD Support (PXOC) (04G-P4-6251-KR)

Hard Drive: ADATA XPG SX6000 Lite 512GB M.2 2280 PCIe Gen3x4 NVMe Read

CPU Cooling: AMD CPU stock fan

OS: Windows 10 Pro 64-Bit (10.0, Build 10240) I have also used newer builds but received the same errors

System 02:

Motherboard: ROG STRIX B550-F GAMING (WI-FI)

BIOS: Version 1004

Processor: AMD Ryzen 9 3900X

Memory: G.SKILL F4-3200C16Q-32GTZR Trident Z RGB Series 32GB, 288-Pin SDRAM DDR4-3200MHz (PC4 25600)

Video Card: ASUS GeForce RTX 2080 O8G Dual-fan OC Edition GDDR6 | 1830 MHz Boost, 14000 MHz Memory | PCIe3.0, 1xHDMI 2.0, 3x DP, 1x USB Type-C | DUAL-RTX2080-O8G

Hard Drive: WD Black SN750 500GB PCIe Gen3 x4 NVMe M.2 2280

OS: Windows 10 Pro 64-Bit (10.0, Build 10240) I have also used newer builds but received the same errors

CPU Cooling: NZXT X63

Power Supply: Corsair HX750 80Plus Platinum

Both systems experience the same issue.

System 02:

• I tested with two different processors (System 01 and 02 processors),
• Two different motherboards (exchanged the original),
• Three different graphics cards (Both NVIDIA and an AMD 5700XT),
• I have removed the M.2 drives and tested using a 2TB Seagate hard drive.
• Two different Corsair 80Plus Platinum power supplies 750w
• Updated the BIOS Settings and default BIOS settings
• Reinstalled multiple versions of Windows 10
• The system was running fine for about three weeks then this happened out of no where
• When I try to install the drivers for the NVIDIA card I receive an error: This NVIDIA graphics driver is not compatible with this version of Windows. Even if I use the NVIDIA experience and have it auto download the drivers.

System 01:

Pretty much the same as above, except it never worked...

Are AMD processors this bad? I haven't used AMD since my Athlon 1600XP, and have had a few Intel systems since then and all were rock solid. This is a horrible experience with AMD and I am looking for help, someone has to have had this issue, I see a lot of post but no one is listing how they solved it.

Asus wants me to send the board back, but they wont pay for shipping on a two month old mobo... So I would not recommend buying Asus at this point because their Canadian customer service sucks.

[mstfbsrn980]

Have you tried with a different brand PSU? I very rarely got this error message when I ran tests by replacing the PSU.

[rlauzon]

I have tested with two brand new Corsair PSU's Platinum+. and an Antec Gold+ certified... Do not suspect it to be the power supply any more.

[mstfbsrn980]

The error message you got indicates that the CPUs are not working stable with very high probability.  Use a multiplier of 42 for 3900x and 41 for 3600x, the problem will most likely improve with BIOS CPU Core Ratio. But you should also seek your consumer rights. You can improve the situation just with what I wrote.

[pokester]

I have several Ryzen systems and have not run into this error before. I found a site that list many possible causes that may give you some ideas on what to check: https://www.xtremerain.com/clock-watchdog-timeout/ 

I did notice on that site that memory corruption can be a cause of that error in looking and your motherboard and processor I did not find the CMK32GX4M2B3200C16 memory you used on that mother board QVL list. So that means the correct timings may not be, being applied. 

Also multiple drivers can conflict with each other. You mention you had several GPU installed. Did you uninstall the drivers and run DDU before installing another GPU?

Also on those PCIe 4.0 boards you might want to try changing the slot to 3.0 if you are using a PCIe 3.0 card like the Nvidia cards. 

I assume you have checked all of this with bios defaults, no PBO or DOCP/XMP?

Also are you using the latest bios for both boards?  for instance the b45m-a board appears to have a newer bios available. You should check the other board too. 

 Version 22032020/08/119.5 MBytes

PRIME B450M-A BIOS 2203
Improve system stability

Also are you running Windows 10 2004 the current version of Windows? And is it fully updated?

Are you loading the latest AMD Chipset driver?

Also 3rd party anti malware software can cause issues. If you loaded anything other than Defender you might try uninstalling it to see if that helps. 

[rlauzon]

I found a site that list many possible causes that may give you some ideas on what to check: https://www.xtremerain.com/clock-watchdog-timeout/ 

• I found this site as well... didn't help much

I did notice on that site that memory corruption can be a cause of that error in looking and your motherboard and processor I did not find the CMK32GX4M2B3200C16 memory you used on that mother board QVL list. So that means the correct timings may not be, being applied.

• For system 01 that is correct, but I confirmed on other sites that people were using the ram that I purchased with no issues.
• For System 01, I am using certified ram

Also multiple drivers can conflict with each other. You mention you had several GPU installed. Did you uninstall the drivers and run DDU before installing another GPU?

• I am doing fresh install of Windows, I then install the AMD chipset drivers and then move on to video card drivers and it fails every time. Doesn't matter if I am fully updated or no updates to Windows, I have tested both
• I wiped the drive clean and started again when I switched from NVIDIA to AMD Radeon

I assume you have checked all of this with bios defaults, no PBO or DOCP/XMP?

• Yes

Are you loading the latest AMD Chipset driver?

• Yes

Also are you using the latest bios for both boards?  for instance the b45m-a board appears to have a newer bios available. You should check the other board too. 

• Yes, I updated both boards, and ROG STRIX B550-F GAMING (WI-FI) I have done that twice as the first board was replaced because Asus support said it was defective.

Also 3rd party anti malware software can cause issues. If you loaded anything other than Defender you might try uninstalling it to see if that helps.

• I don't load any of that stuff on my systems, and all the troubleshooting I have been doing are on clean installs of Windows

[pokester]

Sorry none of that helped. I really wish I had better advice to offer. I can only tell you I have 3 ryzen systems personally all on MSI b450 Tomahawk or TH Max boards. I have had no issues. Do I truly don't think your issue is with the AMD processor but maybe with more to do with the Asus boards for some reason. I can't bad mouth them as I have no experience with them on Ryzen. I only know that mine have just worked since day one. I got a slight amount of performance improvement last year after a couple bios and chipset updates but that is about the only thing that has changed since first installed.

I went with the boards I did because of all the in depth reviews saying these were good stable boards. I know some other B450 boards had issues with ryzen 3xxx not sure about yours. That however doesn't explain your issues with the B550 board too. 

All I can say is work with ASUS, if you are beyond your return period for your boards maybe if you bought with a credit card you have some buyer protection there and can file a claim and go with something else. 

I should mention that I did have my first board die, but I think it was a surge that killed it. Either way MSI did replace it form me. I did have to pay shipping to them but they paid shipping back. Now they actually sent me back a dead board, so the ate the shipping both ways on that one. 

So maybe explain to ASUS you already did one round of returns on this and maybe they will pay the shipping. 

I am sorry you are having all this trouble. While many AMD GPU users have issues, typically Ryzen has been pretty good and your issues to my knowledge are not normal. 

[mstfbsrn980]

Could you follow this ways?
+ Set CPU Core Ratio value to 41-or-40 for 3600x with the BIOS.
+ Disable C-State features for the CPU with the BIOS.

[rlauzon]

Did not work, thanks for the advice.

[rlauzon]

The battle continues...

Good news, I was able to fix the issue for one of the two systems by using a new USB key and creating a new Windows 10 installer using the latest build. Bad news... the other system is defective using the AMD Ryzen 9 3900X has now been confirmed defective by AMD.

After applying the same fix to both systems, the 3900 started generating new errors

A fatal hardware error has occurred.

Reported by component: Processor Core

Error Source: Machine Check Exception

Error Type: Cache Hierarchy Error

Processor APIC ID: 0

The details view of this entry contains further information.

The value for "Processor APIC ID: 0" would change between 0, 8 and 9

The issue I now have, is that AMD wants me to pay for shipping after spending hours troubleshooting this with Asus and other vendors and it ends up being the processor and AMD did ZERO help to troubleshoot, they want me to pay to ship something that started a month and half after the purchase date... This is horrible, horrible customer service. The last AMD system I will ever deal with. I am almost tempted to get this RMA done and then sell the system and go back to Intel.

AMD has the horrible customer service.

[mstfbsrn980]

Go to BIOS for the system you are having problems with and set the CPU Core Ratio to 35. AMD systems set the core ratio multiplier based on the base clock I guess. I think the suggestion I wrote to you before was wrong.