General Discussions

[Ricard0]

Hello!

I built my pc in March with the new AM5 platform and been running beautifully fine until 1/2 weeks ago. When I started getting the "Memory could not be read" errors, in every applications, specially games. Sometimes I can be there for 5 sometimes for 1h. Rarely takes more than 30m when I play games such as The Last of US or Resident Evil 4. If i'm playing World of Warcraft I'm getting error 132 Memory could not be read all the time. Sometimes 3 times in 10 minutes. This happens always inside the game of course.
I also get a lot of BSOD. Mostly memory related ones I believe, but been getting quite different types of them.
All this at stock settings, not even XMP on. Nothing overlocked unless from factory ofc. PC runs fine for desktop tasks, watching videos, browsing even playing CS GO (which is a not so much hardware demanding game) runs absolutely fine.
Temperatures from every single component are fine and within spec nothing to note.
I cannot activate XMP profile or it won't boot. Boots fine without XMP, theres a curious situation, after I try to boot with XMP and don't manage to I need to clear CMOS and after that I need to play around with the ram sticks until I get the first boot and then will boot just fine from that moment on.

Specs:
AMD Ryzen 7600X
B650 Tomahawk WIFI
RAM G.SKILL Flare X5 32GB (2x16GB) DDR5-5600MHz CL36
Radeon 6800XT
Seasonic 850W Gold modular

So I always been running stock since the system was built except for the XMP profile was on. When I started to get these errors and BSOD I tried updating BIOS (3 different ones), reseting Windows, formating Windows, bought and tried a new M2 SSD, bought a new ram kit, tried many different RAM configurations, one stick, both sticks, different slots.
Software wise I tried all I could find when browsing for fixes for almost 2 weeks. Uninstalling games, running the System files scans, many of other potential fixes won't mention because resetting and installing clean Windows would solve most of them.
I tried 4 different GPU driver versions with and without the AMD Adrenalin software.
All other drivers always updated (including chipset).

I started running stress tests, I ran Memtest86 twice on the ram kits and both came out clean.
Ran 3D mark.
Ran Cinebench.
Ran Prime95 where it sometimes failed the large FFT test, always on 960k. "Hardware Failure detected running 960k fft size"

Then I tried OCCT:
OCCT in Power, 3D, VRAM and Memory in AVX2 mode come clean.
I used to get CPU error running large data set but last tests didn't get any.
Where I get errors are running Memory in SSE mode (more than 3000 errors) and Linpack test came out with 11 errors.

By this point I think the culprit is either the Motherboard or the CPU. I cannot be sure, I also cannot be sure about what to conclude on these errors in these tests.
Getting a lot of memory errors and games crashing directly to desktop or freezing indefinitly I believed the RAM sticks would be the problem but I bought new ones and nothing changes.
I started to think something wrong with CPU IMC? I had that stock BIOS since march where the high voltages on B650 could be a thing, wonder is that could have damaged.
I should also say I reseatted every single component in the system including CPU where I did check for possible burnt marks or bent pins on MB but both seem clear.
Could it be PSU? Had 0 problems on OCCT power test where it draws a lot of wattage and I been crashing on games where the load is actually very low.

I've done so many things I'm for certain missing to write a lot of them but at this point I'm lost I don't know what else to do/try.

I would greatly appreacite if anyone could help me out.

[elstaci]

The first thing that came to my mind was defective RAM Modules. But since you ran MEMTEST86 without any errors that ruled out defective RAM but not incompatible RAM.

Found this thread concerning the Linpack error and CPU errors. Seems the fix was reconfiguring the RAM Memory in BIOS: https://linustechtips.com/topic/1493622-occt-linpack-cpu-tests-both-fail-in-a-whea-error/

First is G.SKILL Flare X5 32GB (2x16GB) DDR5-5600MHz CL36 listed on your Motherboard's QVL LIST For your Processor as being compatible?

If it isn't listed then go to G-Skill Memory calculator and see if your RAM is compatible with your Motherboard go to the RAM Part Number QVL List for Motherboards.

If it is listed see if it can be overclocked or not and what is its maximum Ram Speed the Motherboard will support with your current RAM Memory.

Second if it is due to RAM incompatibility just install one RAM Stick in DIMM Slot A2 and see if the problem occurs with XMP enabled and disabled. Sometime incompatible RAM will work with one stick installed but as soon as you install a second Ram stick you start having incompatibility issues.

Then open a MSI and GSKILL and AMD SUPPORT Ticket and ask them their opinion concerning the RAM issues you are having.

AMD Support to rule out a bad CPU causing the Memory issues from here: https://www.amd.com/en/forms/contact-us/support.html

I doubt you have a PSU issue with a Gold rated 850 Watt PSU installed.  If you ran the OCCT PSU Test and it passed without shutting the test or PC down that means your PSU is good enough for your computer.

So either you have a incompatible RAM MEMORY hardware issue or a defective Motherboard or possibly CPU.

NOTE: Also make sure you have the latest Non-BETA BIOS version installed version:

Went to your MSI SUPPORT for your Motherboard Compatibility MEMORY Chart and found only 2 G.SKILL RAM rated at 5600 MHZ as being compatible: https://www.msi.com/Motherboard/MAG-B650-TOMAHAWK-WIFI/support#mem

Those two were only one listed for your MSI Motherboard model.

[Ricard0]

Thank you so much for your reply!

I bought this kit precisely because it was on motherboard's compatibility list as I also thought it was a RAM problem and I wanted to solve or rule it out as a possibility. 

I also do have the BIOS updated to latest version!

I tried both RAM kits, the previous and the one I'm using now in many different configurations, including running with one stick, and unfortunately the crashes and errors persist. 

Also everytime I try to overlock the RAM, wether it be manually or using XMP profile the system won't boot, won't even let me reach BIOS making me go for a CMOS clear.

[elstaci]

Sounds like a hardware issue.

I suggest your open a MSI SUPPORT ticket since are using the RAM from their QVL List and one to AMD SUPPORT in case it might be a CPU problem and finally one to G.SKILL SUPPORT to see if they believe you RAM might be the problem.

Truthfully, I don't believe it is your RAM but a Motherboard or CPU issue but no harm in opening all those tickets.

Please post back the replies to any of the Support tickets you might open up. Thanks.

NOTE: Any errors showing in Windows Device Manager by any chance?

[Ricard0]

Quick update: I managed to start booting just fine with XMP enable, boots just took longer than 1m but with Memory context on it came down to 18s. Never failed to boot, just took long for the memory training to be done. 

I do think it might be a IMC problem in CPU. Is it possible to have the pc running perfect for 5 months and the IMC suddenly go bad? Could those high voltages problems the B650 boards had with the first BIOS versions could have damaged it?

[Ricard0]

I also think it's one of those, possibly CPU. I been running the stock BIOS until I started having these problems couple weeks ago, I did not track voltages as I only learned about that voltage cap was introduced in a later BIOS version and I fear it might have damaged the chip somehow? However I did not see any physical damage signs when I popped the chip out. 

Also device manager does not show any errors, no.

Thank you for your messages! I did open tickets with AMD and MSI already, didn't think to do it with G.SKILL since I currently have 2 ram kits and the problems persist but they do be both from G.SKILL so I also might do it. Will update if I have any news from them.

[Hiramabiff]

There is a new windows 11 update that was released last week that caused this with mainly msi mother boards.  

https://www.pcworld.com/article/2041506/windows-11-update-causes-bluescreen-of-death-microsoft-searc...

update again and remove that update. 

[Ricard0]

Of course silly me, I knew I would forget of something when writing the post.

I'm still on Windows 10, the last update I have similar to that one is KB5029331, but I never got the unsupported processor error. 

[Ricard0]

Update after some more troubleshooting:

A friend sent me his GPU and PSU so I could do more testing and the problem persisted with both. 

However, I tried the CPU iGPU and beside the obvious performance impact, the system is stable with no crashing during games.

So the conclusion is:

- When I run perfectly working discrete GPUs the games and even Windows itself crash.

- When I run the processor iGPU the games and system won't crash.

At this moment the problem can only be either the CPU or the Motherboard.

As I do not know for sure how these two components interact with a discrete GPU it would be incredibly helpful if someone could give me some insight so I can try understand which one is most likely the problematic one.

Thank you!

[elstaci]

Generally, the first PCIe slot #1 on the motherboard closest to the CPU is controlled by the processor. So it could be a defective PCIe slot #1 or the CPU that is controlling the PCIe slot.

If you install the GPU card on a different PCIex16 Slot and it doesn't crash it could be a defective Mobo PCIe slot or the CPU.

Possibly your motherboard's PCIex16 #1 Slot might be damaged or shorting out when you install the GPU card. Remove the GPU card and visually check the PCIe slot's contacts and see if you see any bent or damaged contacts.

Best way to find out is by installing your CPU on another compatible PC or installing another CPU in your PC.

If the GPU card doesn't crash anymore with a different CPU then you probably have a defective CPU but if it continues to crash then you have a defective Mobo.

I would open a AMD SUPPORT - Warranty ticket to see if they believe you need to RMA your APU Processor to be checked or replaced: https://www.amd.com/en/forms/contact-us/support.html

NOTE: In BIOS, when you install a GPU Card it should automatically make your GPU card the Main Display Adapter and your IGPU secondary unless BIOS Setting is specifically set for IGPU. Otherwise generally its default setting is "AUTO".

EDIT: Since you were also getting many Memory crashes that could also be related to the APU processor.

[Ricard0]

Thank you very much for this answer, I tried the GPU in the other PCIE slot and it still crashes so in this case is most likely a motherboard problem is that correct? 

[annaharrow]

RAM Compatibility and Configuration:

• Check if your G.SKILL Flare X5 RAM kit (32GB DDR5-5600MHz CL36) is listed on your motherboard's QVL (Qualified Vendor List) for compatibility with your processor.
• If not listed, consult the G.SKILL Memory Calculator to confirm RAM compatibility with your motherboard.
• Test RAM stability by using one stick at a time in DIMM Slot A2, both with XMP enabled and disabled.  

Support Tickets:

• Open support tickets with MSI, AMD, and G.SKILL to seek their opinions and assistance regarding the RAM and possible CPU issues.
• AMD Support can help diagnose potential CPU problems. 

BIOS Update and Windows Updates:

• Ensure you have the latest non-BETA BIOS version installed for your motherboard.
• Stay updated with Windows 10 (or 11) to avoid compatibility issues.

GPU and PCIe Slot Testing:

• Test GPU stability by installing it in a different PCIe x16 slot, if available, to check if the issue persists.
• Inspect the contacts of the PCIe slot closest to the CPU for any bent or damaged contacts. 

iGPU Testing:

• Running the CPU iGPU without discrete GPUs resulted in system stability. This suggests a potential issue with the CPU or motherboard. 

Hardware Isolation:

• To pinpoint the problem, consider testing your CPU in another compatible PC or trying another CPU in your current system.
• If GPU crashes continue with another CPU, the motherboard may be defective. 

By following these steps, you should be able to identify whether the issue lies with the RAM, CPU, or motherboard and proceed accordingly with troubleshooting or warranty claims.

[Ricard0]

Thank you very much for the answer! I will try to find a way to test my CPU in a different system as I don't think the ram is problematic, its the second kit I tried and it is listed in motherboard's compatibility list.

I did speak with all AMD, MSI and G.SKILL support, unfortunately they didn't been helpful at all. MSI and G.SKILL told me to activate warranty, even though I wasn't even sure which part was problematic. With AMD I had 3 emails with the same troubleshoot ste that I had done multiple times as I wrote in my first ticket to them (uninstall and reinstall GPU drivers), when what I really needed help with is to understand if the problem is the CPU or board. The help, thankfully, I found in here with you, the community. 

So thank you again for that.

[Ricard0]

*Update*

Replacing motherboard and CPU fixed the problem.

Thank you for all the replies and the help!

[elstaci]

Please mark your last reply "Solution" so other Users will know how you fixed your problem.