PC Processors

[mrtasty69]

Components: b450 tomahawk max, 1080 ti asus FE, ballistix 3200mhz c14, 3700x, ssd team group 240gb //// new motherboard: asus prime a520M-K (updated motherboard bios to latest, re installed gpu bios, re installed chipset drivers)

I've been having issues with my pc out of nowhere it began restarting over and over, at first it would let me get into the desktop but after 4 times it started looping the restart in logo screen I thought it was the ssd so I reinstalled windows and it started restarting again, reinstalled in a hdd and it worked but when i tried installing my nvidia drivers it wouldn’t use my card (wouldn’t load nvidia control panel) it tried to tho, it did the restart of display as to trying to recognize the card and run the display through the driver, but went back to displaying image without using the driver (tried re installing in 3 different hard drives, usb one, ssd, and hdd and in some of them i would get the blue screen nvidia nvlddmkm.sys after installing the os) I got it to boot in a few of the reinstalls so i went to device manager and it said under my gpu (1080 ti fe asus) code 43 which is usually a hardware problem I thought it was my pcie 3.0 port that was damaged.

I switched to my secondary pcie2 16x (4x lanes) on the tomahawk and the gpu worked, accepted the driver and showed everything working in gpu-z.

Never had any issues with any of the components before, no weird temperatures and everything has been stable for almost a year since i got the rig, it’s still stable and I’m able to use it but not through the pcie 3.0 port that communicates directly with the CPU.

What I've done:

• Replaced mobo to discard the pcie 3.0 slot, displayed the same error 43 so it's not the mobo.
• Same ram sticks have been running flawless throughout all of this including right now, ran the cmd ram check up and it didn't detect anything.
• Installed my GPU in another rig (pcie 2.0 16x port) intel and it works flawless without errors.
• It's not my storage as I'm using it rn with my 1080 ti in the x4 pcie 2 slot.
• Tried running a 730gt in the pcie 3.0 slot of my tomahawk and it crashed with the same blue screen error (nvlddmkm.sys)
• It's not my PSU as the 730gt does not use any power from the PSU directly and it still crashed (the 730gt did work in the pcie 2 4x of the tomahawk) (I also went in the bios and tried in pcie settings gen 1, gen 2 and gen 3).
• Reset CMOS, reinstalled chipset drivers and removed nvidia drivers with DDU multiple times.

It seems that the restarts and crashes were caused by the CPU itself because the CPU and GPU weren’t able to communicate and the system was trying to force the os to run through the nvidia driver... maybe the card overloads the cpu?

Is there any way to fix this without replacing the CPU?

I really don’t understand how can all of my components work and run great but not together with the CPU?

Also what exactly is wrong with my CPU  if it doesn’t overheat or crashes and its stable even if I overclock it to 4.2 but can't display graphics...

Images:

Blue screen error I got when trying to boot through nvidia driversmy gpu running in pci 3.0 in sisters rig

[elstaci]

I also have a Ryzen 7 3700X with a Asus GTX 1070 on a Asus Motherboard without any issues.

One thing I noticed in your image of Windows CPU. It only shows 4 cores with 4 logical cores. It should show 8 Cores with 16 logical cores like mine does:

But I noticed you have an Intel Processor I5-3470 installed and not the Ryzen 3700X.

So please show the same image but with your Ryzen CPU and not the Intel CPU.

Are you having the same issue with the PCIe 3.0 1st PCIe slot with both the Intel and Ryzen CPUs?

Is your RAM Memory listed on your Asus Motherboard's QVL LIst for the 3000 series processors?

These are the only Crucial RAM Memory listed for your Asus  Motherboard: https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-A520M-K/HelpDesk_QVL_Memory/

[mrtasty69]

The gpu image says its in my sisters rig and the point of that is to discard the gpu as the problem, that is not my setup...

the ram or the compatibility on the second motherboard has nothing to do with this bro, also it is compatible.

I answer all of your questions in the initial post

heres my rig, 1080ti in the pcie 2.0 4x slot, I didnt include it cause i didnt see the point 

[elstaci]

then open a AMD Warranty request and see if AMD believes your CPU needs to be RMAed to be checked if it is defective or not.

[ryzen_type_r]

Yeah it does sound like there's something wrong with the PCIe lanes from the CPU.

If that were my PC, I'd also try some other non-Nvidia PCIe device in that first slot just to see if they also have issues being recognized.  Could be any PCIe device, like a sound card, RAID card, even an AMD video card.

I'd also try reseating the CPU.  What is the SoC voltage at?  I'd try bumping it up a tiny bit, don't go more than 1.1V total though.

Failing all of that, I'd say it's RMA time.

[mrtasty69]

I will do what you adviced, if the problem is the lanes from the CPU does that mean the CPU is degrading in anyway or its just an isolated issue with it because family could use it for non gaming.

Also do you have any idea as to why would the lanes from the CPU break like that?

thanks

[ryzen_type_r]

Don't know, it's a very unusual problem that I personally haven't seen before.

You can certainly continue to use the CPU as it is, but if for example if the SoC silicon is degrading, who's to say what will start failing next.  Could be memory, or USB (as the SoC manages all of these).  So you might as well RMA it now while it's still under warranty.

[mrtasty69]

why would the silicon degrade so much when I've never had temp issues and the CPU has exactly 1 year? just a bad unit?

[vialli100]

Do you have a spare PSU or can you borrow one?

It would have to be a decent make..

Just a thought..