Processors

[SamKO]

Hi everyone,

I have two systems:
PC1: Win10, ASUS Rog Strix x570-E, 3900x, RTX 2070S, Corsair 750x, G.Skill TZ Neo 32GB DDR4 3200MHz(2x 16GB)
PC2: Win10, ASUS Rog Strix x570-E, 3900x, Rx 480 8GB, Corsair 750i, HyperX Fury Black 30GB DDR4 3200MHz (4x8GB)

I got a strange issue with my Razor 9 3900x processor in PC2.

It started with the PC2 just out of the blue shooting down and after that it went in a restart loop.
After all sorts of debugging which includes:
Installing a new system on a new SSD
Swapping out GPUs
Swapping out PSUs
Swapping out RAM
The issue turned out to be the CPU.

The issue can be explained as follow, if I connect any GPU to any of the two PCIe 4.0/3.0 x16 slots on the ROX x570-E motherboard (https://dlcdnets.asus.com/pub/ASUS/mb/SocketAM4/ROG_STRIX_X570-E_GAMING/E15826_ROG_STRIX_X570-E_GAMI...) with the CPU that has the issue, the PC will go into a restart loop where it will restart the pc just before the Login screen of Windows 10.

This can be replicated on both motherboards with a RX 480 and a RTX 2070S. The system wont go into a restart loop if the Graphic driver isn't installed(using windows generic driver),but the second it installs the driver the PC instant Shuts down and restarts and goes to restart loop. Swapping the CPU with the working 3900x fixes this and the system works fine.

Current workaround for the affected CPU is to connect my GPU to the 3rd slot available at the bottom which is a PCIe 4.0 x16_3 slot, then the issue is gone, and the PC works normally.

Anyone got some ides what happened there and if this can be somehow fixed?

Thanks.

[elstaci]

From what you explained that if you install any GPU card in PCIex16 slots 1 or 2 with one processor in any of the 2 motherboards,  it goes into a constant start loop but when you install the second processor everything works fine.

Does sound like the processor went bad for some reason. They generally are very reliable and sturdy.

PCIex16 Slots 1 & 2 are controlled by the CPU and not the CHIPSET drivers.

If it was working fine in the past and this issue just started very recently than it looks like the CPU is defective.

You have a 3 year AMD Warranty if you (Original Owner) purchased the processors from a Retailer in a sealed official AMD Retail box.

I would open a AMD Warranty ticket and see if they recommend to RMA your bad processor as long as the Retailer's Warranty or Return date is expired  from here: https://www.amd.com/en/support/kb/warranty-information/rma-form

[SamKO]

I did purchased the processors from a Retailer in a sealed official AMD Retail box, but the local retailer gave only a 2 year warranty, which expired 2 months ago so I'm probably out of luck there.

Well, it's still usable for now, both M.2 NVMe slots work without any issues, the RX 480 8GB connected to the last(bottom - 3rd full length PCIe) PCIe port works as well and the performance isn't noticeably impacted by that. So I can live with it till it's time to switch to DDR5 and the latest AMD processor.

[ryzen_type_r]

Very interesting issue.  I'd try bumping up the SoC voltage on that CPU to see if it can be stabilized.  Default is 1.000V, try going up in .01V increments.

You can also try playing around with the PCIe settings for the CPU's PCIe controller.  Maybe try x8 and see if that works.

[elstaci]

if @ryzen_type_r suggestion doesn't work you still have a year left of AMD Warranty on your processor. So you can still open a AMD Warranty request.

You can also open first a AMD SUPPORT Ticket under "Warranty" and ask them if they believe you should RMA the processor from here: https://www.amd.com/en/support/contact-email-form

[SamKO]

Thanks for the suggestions, will wait for the weekend to try some of the things as I need the PC right now for work.
I will make sure to report back if any of the possible solutions fixed the issue.