PC Processors

[AustinML]

I recently built a new PC after getting lucky in a best buy drop and snagging a 5800x and 3070. It includes a few random parts I've accumulated over the years which I added to the pcpartpicker list here:

Type Item Price

CPU	AMD Ryzen 7 5800X 3.8 GHz 8-Core Processor	-
CPU Cooler	Cooler Master Hyper 212 Black Edition 42 CFM CPU Cooler	$33.99 @ Amazon
Motherboard	MSI MPG X570 GAMING PRO CARBON WIFI ATX AM4 Motherboard	$258.02 @ Amazon
Memory	Corsair Vengeance LPX 32 GB (2 x 16 GB) DDR4-3200 CL16 Memory	$149.99 @ Corsair
Storage	Samsung 970 Evo 500 GB M.2-2280 NVME Solid State Drive	$64.98 @ Amazon
Storage	Samsung 860 Evo 1 TB 2.5" Solid State Drive	$109.99 @ Adorama
Storage	Seagate Barracuda 3 TB 3.5" 7200RPM Internal Hard Drive	$175.00 @ Amazon
Video Card	Gigabyte GeForce RTX 3070 8 GB GAMING OC Video Card	$1099.99 @ Amazon
Case	Fractal Design Meshify C ATX Mid Tower Case	$100.98 @ Newegg
Power Supply	EVGA G3 750 W 80+ Gold Certified Fully Modular ATX Power Supply	$139.90 @ Amazon
Operating System	Microsoft Windows 10 Pro OEM 64-bit	$139.88 @ Other World Computing
Monitor	Dell S2716DG 27.0" 2560x1440 144 Hz Monitor	$700.00 @ Amazon
Monitor	Dell U2417H 24.0" 1920x1080 60 Hz Monitor	$279.95 @ B&H
Prices include shipping, taxes, rebates, and discounts
Total	$3252.67
Generated by PCPartPicker 2021-01-10 13:11 EST-0500

Also attached are a Logitech G19 keyboard, G600 mouse, and a Microsoft Xbox One wireless adapter for a gamepad.

Initially this computer worked great and I actually played through all of Cyberpunk 2077 with it. However, after doing so and playing other random games (A Monster's Expedition, Cloudpunk, Medal of Honor Above and Beyond, Gears 5, etc.) I would get BSOD. I did what I could to analyze these BSODs using Event Viewer and Bluescreenview but my pc stopped generating dump files on Jan 5. The last dump file showed errors in ntoskrnl.exe and PSHED.dll but NOT hal.dll as I saw in several other similar troubleshooting posts. One issue I have in troubleshooting is that I don't have a reliable way to generate a BSOD - no stress test I have run has caused a BSOD. However, what will eventually cause it is gaming - usually within the first 10-20 minutes. Here is what I've done so far to try to narrow this down:

• Run MANY stress tests including memtestx86, Prime95, 3dmark, cinebench - all with no issues and pretty high temps (90 on the Ryzen, 88 on the GPU, 65-70 on motherboard stuff).

• Swapped out the 3070 for my old 1070 and 970's

• Swapped out the memory for some older DDR4-2400

• Tried available BIOSs that work with the 5800x - I've tried messing with OC settings but honestly not super confident in how to do that

• ran SFC and DISM to check Windows installation

• removed any and all drivers and reinstalled one by one

• Tried older nVidia drivers

None of this helped and I'm worried some step in there actually aided in me not getting dump files anymore. Here is what I have not yet tried:

• Swapping out power supply, cpu, or motherboard

• Complete re-install of Windows - might try this afternoon but highly skeptical this will do anything

Random thoughts:

• The boot drive is the 500gb NVMe drive that sits pretty much right under the GPU. Is it or it's controller getting too hot? I try to watch the temps in HWMonitor but nothing seems to get beyond ~75C. I may try just moving it to the other NVMe slot but seems like the curent slot should be fine if they designed that way (i.e. putting the fast 4.0 slot under the primary GPU slot)

• Maybe swap out my old 650W power supply for this 750W. Though despite all the calculators saying the 750W is plenty for my setup even with overclocking maybe I should just return the 750W while I still can and get an 850W.

• Adding in more fans - I only have one intake and one "extake" along with the cpu fan. The system feels pretty warm after it's crashed but as I said, none of the actual monitored temps seem outside of the norm.

• Bite the bullet and try to exchange the proc and motherboard. I'm worried I got lucky with the proc and won't be able to get another for months.

Any thoughts you have here about other things I can try or anything obviously wrong with the build would be great. It's SUPER frustrating as this PC is amazing for one of the things I built it for (editing 4k video footage) but can't game reliably. I know these types of BSODs are hard to nail down but I'm willing to try just about anything at this point to save this build. I've read through a LOT of posts here and it seems like others are having similar problems but not this exact problem. Could this be caused by some of the other voltage/overclocking issues people are having?

Thanks in advance for your help!

[elstaci]

Try using OCCT to stress test your PSU and CPU and GPU. See if it crashes during the PSU Test. Keep an eye on Temperatures and PSU Outputs.

If it doesn't crash while under several Stress test then it is possibly you have a driver issue.

Try running your Windows in a Clean Windows Desktop and see if it BSODs while gaming. Basically disabling all 3rd party startups and drivers except Microsoft.

Also enable Driver Verifier. This will let you know which driver is causing the BSODs. You will need to read how to use it unless you are familiar with it already.

Run DXDIAG.exe and save the DXDIAG file. Open the file and go to the last category "Diagnostics". That will show you all the files that are having issues on your computer.

[AustinML]

Thank you for the quick response! Here is what I found after running through the suggested steps:

• OCCT showed no errors and I didn't see anything out of the ordinary in any of the graphs/results. Though I am far from an expert when it comes to "normal" behavior with voltages. No crashes.
• I also ran Geekbench as it was suggested in another thread as more "real world". No crashes there either.
• Disables any and all startup stuff outside of MSFT - even unplugged all USB devices. The one thing that I can always do to get my system to crash is run Medal of Honor Above and Beyond. It is a VR game that I remote desktop into the PC with my Quest 2 headset. I also get crashes playing much less system-stressing games like 2D indies. Strangely, Cyperpunk 2077 doesn't seem to crash the system despite seemingly stressing it.
• Ran Ryzen Master. Enabling Precision Boost Overdrive tested fine with no crashes. However, when I tested "Auto Overclocking" it crashed almost immediately. This seemed to be OK though as on reboot, Ryzen Master loaded up again, identified the previous crash, and disabled the Auto Overclock.
• Out of curiosity and since my old system also had an EVGA Power Supply, I swapped my new one into the old system to see if it had any issues. It did not running all of the same games.
• Driver verifier didn't show any errors but not entirely sure I set it up right. I attempted to research it a bit but there was a lot of conflicting information about it's usefulness in Windows 10.
• Ran DXDIAG.EXE and pasted the Diagnostics info below. No errors on running it but it did generate stuff in the Diagnostics area of the saved file.

After digging through this forum a bit it seems my issue is similar to what others are seeing around the power states of this processor. I haven't found any situations that seem to be exactly mine but it seems many are having issues with the power management even with the ComboAM4PIV2 1.1.9.0 update. I attempted some manual settings around the power management found in another post (Solved: Ryzen 5900x: System constantly crashing/restarting... - Page 30 - AMD Community) but that made games crash almost immediately on loading into them.

So at this point I think I'm either having power issues with the 5800x similar to what others are seeing or there is some very odd nVidia driver issue that is not playing nice with the AMD proc or chipset (as the 3070 runs fine in my old Intel system). I wish there was some test I could run that would help me definitively say it is absolutely the processor, motherboard, or odd nvidia driver issue. I worry a warranty replacement of the processor is my best bet at this point but given supply issues I don't want to waste anyone's time or be without my system for an extended period of time. Are there other tests or things I can try to help narrow down the issue further? It'd be great if I had a reliable way to cause the crash to so I could test more easily.

DXDIAG.EXE output (Diagnostics):

---------------
Diagnostics
---------------

Windows Error Reporting:
+++ WER0 +++:
Fault bucket 1244664781714133073, type 5
Event Name: RADAR_PRE_LEAK_64
Response: Not available
Cab Id: 0

Problem signature:
P1: CpuOcct64.exe
P2: 0.0.0.0
P3: 10.0.19042.2.0.0
P4:
P5:
P6:
P7:
P8:
P9:
P10:

+++ WER1 +++:
Fault bucket , type 0
Event Name: BlueScreen
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffbe859fa03028
P4: bc000800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER2 +++:
Fault bucket , type 0
Event Name: BlueScreen
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffbe859fa03028
P4: bc000800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER3 +++:
Fault bucket , type 0
Event Name: BlueScreen
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffe48e5fae3028
P4: bc000800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER4 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffb88f7fceab90
P4: faa00000
P5: 80b
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER5 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffff800be0025420
P4: fc000800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER6 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffbf04400b24c0
P4: fc000800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER7 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffc809f42568c0
P4: faa00000
P5: 80b
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER8 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffffe30954a458c0
P4: fe800800
P5: 1010135
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

+++ WER9 +++:
Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 124
P2: 0
P3: ffff90835d2bea80
P4: baa00000
P5: 2010b
P6: 10_0_19042
P7: 0_0
P8: 256_1
P9:
P10:

[elstaci]

Since the 5000 series CPU are so new I have very little information accumulated so far. But sounds like you are on the right path as far as being a Power issue with the CPU.

You can open a AMD Service Request (Official AMD SUPPORT) under "Warranty" and see if AMD believes you should RMA your processor to be checked or replaced unless it is still under Warranty with the Retailer's Return date: https://www.amd.com/en/support/contact-email-form

It is possible you have a defective CPU.

By the way, When the 3000 series CPUs came out AMD added a BIOS setting concerning the PSU called "Power Supply Idle Control".  The setting is for low or normal current.  If the 5000 Series has the same settings in BIOS see if it is set for low and if it is change it to Normal Current and see if it continues to BSOD.

This BIOS Setting when set to low will cause some PCs to shutdown during idle or under light loads because the PSU can't supply the proper voltage when it is set on Low.

[JoltCola]

Since it only happens under load it doesn't sound like the WHEA reboots many have been seeing, but give a BIOS with AGESA 1.1.9.0 a shot, maybe that'll help.

Otherwise, my guess is it's your PSU or your CPU needs a bit more voltage to be stable under load.

[AustinML]

I am actually running the new AGESA 1.1.9.0 BIOS now but it didn't seem to help - meaning I still get BSODs when gaming but I'm not sure if they are any more or less often.

There are so many BIOS setting I'm not sure which would be able to increase the voltage a bit but I'll research that for my BIOS. So far though anything I change from Auto to a profile or manual setting definitely makes things worse.

[AustinML]

Interesting - I'll try this. It is very strange that I only get the BSOD during gaming and not any other operation like browsing, photo/video editing, etc. or any stress test I've run. Makes me wonder if gaming is some sort of right in the middle use case where it uses multiple cores at different states rather than a single core issue or all core issue which a stress test might now.

[AustinML]

OK - I adjusted the Power Supply setting but unfortunately still saw crashes (Gears 5 for this test) about 10 minutes into gaming.

[AustinML]

I believe I finally may have resolved this issue (fingers crossed). I think I should have been a bit more methodical in my troubleshooting. Long story short, I think it was ultimately the very thing I first thought it was - a driver issue.

With my 3070 I had uninstalled the driver in just the standard windows 10 way. In later troubleshooting steps I had used the DDU from guru3d to remove the driver and reinstall it. However, when doing so I had either my 1070 installed or my even older 970 installed. Totally my spitballing theory but I think something about installing the nvidia driver with an older card installed in my setup causes issues when I replace it with my 3070. That or I just got plain unlucky in adding/removing cards and installing/uninstalling drivers corrupting some portion of the driver that causes BSODs with no dump files early into starting up and playing a game but not in any other tested usage scenario or stress test. One final difference is I was able to install the display driver without installing the GeForce Experience - no idea if that helped.

So hopefully I now have this issue resolved and with what all other tests show to be a good processor. Thanks for your help here and I will update further if I start to see issues again.

P.S. If you suspect a display driver issue, definitely try DDU to completely remove the old driver from your system!

[seikros]

Hello! ,Your Solunion  finally still working? 

I have the same issues but i didnt install a new gpu   just changed my mb x470 and 2700x to pro carbon x570 and 5800X.