Question PC randomly restarts while gaming after swapping my motherboard, CPU and PSU

[Gaff1]

You weren't running the "large FFTs" test though. The Large FFTs test "stresses memory controller and RAM". A half hour isn't really much of a test either. Several hours would be more conclusive. I test for at least 3 to 4 hours.
As KingLoki suggested try running HWiNFO64 and getting a screen capture of what it reports for your voltages, CPU speeds, FCLK, memory clocks, CPU/DRAM voltages then posting it here. You might have to take two screen caps to get everything.

im pretty sure i did run large ffts, also i dont have time to leave it running for 3 to 4 hours, max would be 1 hour.
i will do it again tomorrow with hwinfo

 

[80251]

Gaff1 your output didn't look like the output from the LargeFFTs Prime95 test. I've got an older version though. The test will literally use 95% of your RAM. The below is what it should look like:

 

[Gaff1]

Gaff1 your output didn't look like the output from the LargeFFTs Prime95 test. I've got an older version though. The test will literally use 95% of your RAM. The below is what it should look like:

i ran the test again this time for 1 hour, here are all the screenshots

 

[80251]

I'm not an AMD Ryzen expert at all but are you using DDR5? Because the timings for your RAM look way off for DDR4. I've never seen DDR4 using 22-22-22-52 primary timings or a tRFC of 1700 ever, even up to 4522Mhz (but that was on an Intel Comet Lake build).

You're not overclocking your video card/GPU are you?

Have you scanned your system for viruses?

You're going to need an AMD Ryzen 9 expert to look at your HWiNFO64 stats.

 

[Gaff1]

I'm not an AMD Ryzen expert at all but are you using DDR5? Because the timings for your RAM look way off for DDR4. I've never seen DDR4 using 22-22-22-52 primary timings or a tRFC of 1700 ever, even up to 4522Mhz (but that was on an Intel Comet Lake build).

You're not overclocking your video card/GPU are you?

Have you scanned your system for viruses?

You're going to need an AMD Ryzen 9 expert to look at your HWiNFO64 stats.

im using ddr4 3200mhz.

im not overclocking my gpu.

im on a fresh install of windows 10 so no viruses.

 

[80251]

There's no way the timings should be that loose for DDR4 3200. For me, too loose timings were as problematic as too tight timings, but I'm on Intel.
What RAM are you using? What is the specific type and manufacturer?
HWiNFO64 is showing you what your RAM timings should be, but only up to 2666Mhz. Maybe your RAM isn't rated to run at 3200Mhz. at all. I'd try downclocking your RAM to 2666Mhz. and setting the timings according to what HWiNFO64 is reporting should be set for 2666Mhz.
You should also check the QVL list at asrock for your motherboard, it could be it doesn't support 32GB RAM sticks or only at slower speeds like 2666Mhz.

 

[Gaff1]

There's no way the timings should be that loose for DDR4 3200. For me, too loose timings were as problematic as too tight timings, but I'm on Intel.
What RAM are you using? What is the specific type and manufacturer?
HWiNFO64 is showing you what your RAM timings should be, but only up to 2666Mhz. Maybe your RAM isn't rated to run at 3200Mhz. at all. I'd try downclocking your RAM to 2666Mhz. and setting the timings according to what HWiNFO64 is reporting should be set for 2666Mhz.
You should also check the QVL list at asrock for your motherboard, it could be it doesn't support 32GB RAM sticks or only at slower speeds like 2666Mhz.

its sort of a generic ram, brand is ktrok, green pcb
it says on the sticker that is 3200mhz, i found an image of it online so you can se

i was using this stick on my other pc and it had no issues that i remember

i tried using 4 sticks of 8gb 2666mhz and it was also restarting, but now that you said that, i think with those 4 sticks the pc was actually blue screening, and with this ram it just restarts but shows the "same" error in event viewer

 

[80251]

Here's the QVL list for your Asrock x570 Steel Legend and specific CPU (which is a Vermeer):
https://www.asrock.com/MB/AMD/X570 Steel Legend/index.us.asp#MemoryVM

You'll notice the fastest 32GB stick of RAM they tested with was only running at 3000Mhz. and most of the 32GB sticks they tested at 2666Mhz and those were all for well known RAM manufacturers.

A 32GB stick of DDR4 RAM will be at least dual rank and maybe even quad rank, either of which will limit the highest operating frequency. As far as performance goes, you should go back to your 4x8GB DIMM config because now you're only using one channel of your IMC instead of both (HWiNFO64 reports channels as being 1 in the memory timings tab which is not good for performance). 8GB DIMMs are much easier to run at higher frequency than 16GB or 32GB.

I think you should go back to using 4 x 8GB sticks because you would get better performance.

HWiNFO64 also indicates the highest operating frequency for your RAM is 2666Mhz. NOT 3200Mhz (look at the motherboard tab of HWiNFO64). Try reducing the RAM frequency to 2666Mhz and use the timings reported in the motherboard tab of HWiNFO64 for 1333 (which is clock doubled to get you to 2666Mhz.).

I still think your RAM is the issue, especially considering you're using a no-name brand that has a 3200Mhz. sticker but the SPD chip states the fastest speed is 2666Mhz.

 

[Gaff1]

Here's the QVL list for your Asrock x570 Steel Legend and specific CPU (which is a Vermeer):
https://www.asrock.com/MB/AMD/X570 Steel Legend/index.us.asp#MemoryVM

You'll notice the fastest 32GB stick of RAM they tested with was only running at 3000Mhz. and most of the 32GB sticks they tested at 2666Mhz and those were all for well known RAM manufacturers.

A 32GB stick of DDR4 RAM will be at least dual rank and maybe even quad rank, either of which will limit the highest operating frequency. As far as performance goes, you should go back to your 4x8GB DIMM config because now you're only using one channel of your IMC instead of both (HWiNFO64 reports channels as being 1 in the memory timings tab which is not good for performance). 8GB DIMMs are much easier to run at higher frequency than 16GB or 32GB.

I think you should go back to using 4 x 8GB sticks because you would get better performance.

HWiNFO64 also indicates the highest operating frequency for your RAM is 2666Mhz. NOT 3200Mhz (look at the motherboard tab of HWiNFO64). Try reducing the RAM frequency to 2666Mhz and use the timings reported in the motherboard tab of HWiNFO64 for 1333 (which is clock doubled to get you to 2666Mhz.).

I still think your RAM is the issue, especially considering you're using a no-name brand that has a 3200Mhz. sticker but the SPD chip states the fastest speed is 2666Mhz.

i lowered the frequency to 2666mhz and changed the timings, i will do some tests later today and report back

 

[Gaff1]

Here's the QVL list for your Asrock x570 Steel Legend and specific CPU (which is a Vermeer):
https://www.asrock.com/MB/AMD/X570 Steel Legend/index.us.asp#MemoryVM

You'll notice the fastest 32GB stick of RAM they tested with was only running at 3000Mhz. and most of the 32GB sticks they tested at 2666Mhz and those were all for well known RAM manufacturers.

A 32GB stick of DDR4 RAM will be at least dual rank and maybe even quad rank, either of which will limit the highest operating frequency. As far as performance goes, you should go back to your 4x8GB DIMM config because now you're only using one channel of your IMC instead of both (HWiNFO64 reports channels as being 1 in the memory timings tab which is not good for performance). 8GB DIMMs are much easier to run at higher frequency than 16GB or 32GB.

I think you should go back to using 4 x 8GB sticks because you would get better performance.

HWiNFO64 also indicates the highest operating frequency for your RAM is 2666Mhz. NOT 3200Mhz (look at the motherboard tab of HWiNFO64). Try reducing the RAM frequency to 2666Mhz and use the timings reported in the motherboard tab of HWiNFO64 for 1333 (which is clock doubled to get you to 2666Mhz.).

I still think your RAM is the issue, especially considering you're using a no-name brand that has a 3200Mhz. sticker but the SPD chip states the fastest speed is 2666Mhz.

i lowered the frequency to 2666mhz and changed the timings, i will do some tests later today and report back

this seems to have fixed the issue, no more random reboots while under load. thanks for helping

 

[beyondlogic]

Hello everyone,

I'm encountering a frustrating issue with my PC and I'm hoping someone can help me troubleshoot. A few weeks ago, I swapped my CPU, motherboard, and PSU, and since then, I've been getting constant blue screens with the error code WHEA_UNCORRECTABLE_ERROR. The issue started happening even when I was just on the desktop, but it was much more frequent when I played games.

What I've Tried:​

1. Disabled PBO in BIOS: This stopped about 99% of the blue screens. The crashes are now much less frequent, but I still get them in VR.
2. Swapping RAM: I tried using different RAM sticks, but the crashes still occurred, so I don’t think it’s a ram issue.

Current Situation:​

• Now, the system only crashes when I'm in VR. It typically happens a few minutes into the game, and instead of showing the blue screen, the PC just restarts. In the Event Viewer, I get the WHEA error logs, but I’m not sure if I'm just not seeing the blue screen anymore because the PC restarts so quickly. It's possible that the blue screen appears only for a split second before the system restarts.
• Event Viewer Error: The WHEA error shows up with the following details:
  • Reported by component: Processor Core
  • Error Source: Machine Check Exception
  • Error Type: Bus/Interconnect Error
  • Processor APIC ID: 10 (but this number changes with each blue screen)

What I Suspect:​

• CPU Overheating: Currently I only have a 120mm AIO, and I'm starting to wonder if the CPU might be overheating, especially since it blue screens consistently while VR (which is quite demanding). However, I'm not sure if overheating would specifically cause a WHEA_UNCORRECTABLE_ERROR, or if it would just cause thermal throttling and make the game lag.

System Specs:​

• CPU: Ryzen 9 5900x
• Motherboard: ASRock X570 Steel Legend
• PSU: 850W Thermaltake Smart Pro RGB
• RAM: Currently 1 stick of 32gb 3200mhz, but I also tried with 4 sticks of 8gb 2666mhz
• Cooling: 120mm AIO (updated check the end)
• GPU: Gigabyte Radeon RX 6600 8gb
• OS: Windows 10

Alright, need to update some stuff​

• I've upgraded my AIO for a 360mm one

The pc doesn't blue screen anymore, it just suddenly restarts when I'm gaming, but what is weird is that it no longer gives the "WHEA-Logger" error on event viewer, just "Kernel-Power"

Edit: Nevermind, WHEA errors are back

• Could it be a failing PSU?

Here is a screenshot of event viewer, the most recent one was today at 9:50 PM but the last WHEA-Logger was yesterday at 11:44 PM
(every AM its actually PM, the clock was wrong, just the last 2 that correctly says PM, because thats when I fixed)

update your bios to latest version
also you should never buy used psu especially one as old as that i would replace it.

how is the aio placed the new one ?

 

[Gaff1]

Here's the QVL list for your Asrock x570 Steel Legend and specific CPU (which is a Vermeer):
https://www.asrock.com/MB/AMD/X570 Steel Legend/index.us.asp#MemoryVM

You'll notice the fastest 32GB stick of RAM they tested with was only running at 3000Mhz. and most of the 32GB sticks they tested at 2666Mhz and those were all for well known RAM manufacturers.

A 32GB stick of DDR4 RAM will be at least dual rank and maybe even quad rank, either of which will limit the highest operating frequency. As far as performance goes, you should go back to your 4x8GB DIMM config because now you're only using one channel of your IMC instead of both (HWiNFO64 reports channels as being 1 in the memory timings tab which is not good for performance). 8GB DIMMs are much easier to run at higher frequency than 16GB or 32GB.

I think you should go back to using 4 x 8GB sticks because you would get better performance.

HWiNFO64 also indicates the highest operating frequency for your RAM is 2666Mhz. NOT 3200Mhz (look at the motherboard tab of HWiNFO64). Try reducing the RAM frequency to 2666Mhz and use the timings reported in the motherboard tab of HWiNFO64 for 1333 (which is clock doubled to get you to 2666Mhz.).

I still think your RAM is the issue, especially considering you're using a no-name brand that has a 3200Mhz. sticker but the SPD chip states the fastest speed is 2666Mhz.

okay, so i don't know anymore, after changing the timings and frequency of my ram, i tested it on vr which was consistently causing reboots, and it didnt reboot at all, but today when playing another game it happened again, with the same event viewer error

update your bios to latest version
also you should never buy used psu especially one as old as that i would replace it.

how is the aio placed the new one ?

since the pc randomly restarts i dont want to update the bios yet, because if for some reason it happens while im updating then its over

and the psu is not old, they bought it around 2020, the one i had on my other pc is even older, i bought it in 2017

im not sure what you meant, if you're asking where the aio is placed, the radiator is on top

 

[80251]

@Gaff1
Have you tried loading the XMP profiles for your RAM?

 

[Kona45primo]

That's a tough spot. The only whea errors I've had were from to much negative offset doing the core optimizer thing.

Assuming power profile is set to balanced? Guessing it would be.

Might want to try a positive+15 offset just to see if that stabilizes things enough for a bios update.

 

[Gaff1]

@Gaff1
Have you tried loading the XMP profiles for your RAM?

there is no xpm option in the bios, im on version P3.00

That's a tough spot. The only whea errors I've had were from to much negative offset doing the core optimizer thing.

Assuming power profile is set to balanced? Guessing it would be.

Might want to try a positive+15 offset just to see if that stabilizes things enough for a bios update.

i dont remember seeing anything about power profiles in the bios

its not unstable, only happens while gaming, but still random, so it could happen during the update