Question BSOD on sppsvc.exe

[luckzada]

I have several BSOD that points to sppsvc.exe, but i dont know if is a hardware issue or a software error, i noticed that if i lower my base clock (BCLK Frequency) from 100.0 to 96.0 the BSOD become way less frequent.

Link to dump files: https://www.mediafire.com/folder/bjsccu08nqi42/dump

Specs:
CPU: i7 13700k (Stock);
Motherboard: Asus ROG Strix B660-F GAMING WI-FI with latest BIOS (3010);
RAM: 2x8GB Kingston DDR5 6000MHz (KF560C40BBAK2-16) (A_2 and B_2);
GPU: MSI RTX 3060 Ti;
Cooler: Corsair H150i Elite Cappelix;
SSD 1: Kingston FURY Renegade PCIe 4.0 NVMe M.2 1 TB (Windows 11);
SSD 2: XPG S41 TUF, M.2 PCIe NVME 256 GB (Linux);
PSU: XPG Core Reactor 850w;

 

[ubuysa]

Hello, and welcome to the forum!

There are some indications from your dumps that the problem might be RAM related. I'd suggest that you first run the RAM at stock frequency (4800MHz) for a while and see whether it's stable then.

 

[luckzada]

Hello, and welcome to the forum!

There are some indications from your dumps that the problem might be RAM related. I'd suggest that you first run the RAM at stock frequency (4800MHz) for a while and see whether it's stable then.

Thank you!

I tried running at XMP disabled (4800MHz) and with the 4800MHz profile enabled as well, at 100 BCLK it was still unstable. It seems that the frequency dont matter that much, is more about some other value that the BCLK changes with it, because if i set to 6000MHz but with 96 BCLK the system is stable.

Also, i tested the RAMs in memtest86+ on the bios and returned 0 errors.

 

[ubuysa]

Personally I prefer Memtest86 (free) but no RAM tester can ever find 100% of RAM problems. However, let's assume your RAM is OK for now. That changing the BCLK changes the symptoms does indicate that this is almost certainly hardware related. I would next suggest that you restore all clocks and voltages to stock, and then download and run Prime95.

In case you've never run Prime95 before, what it does is stress your CPU, and to some extent your RAM as well. It WILL make your CPU run hot, so you also need to run a temperature monitor (like CoreTemp) to keep a close eye on CPU temperature.

Run all three Prime95 tests (small FFTs, large FFTs, and Blend) for at least an hour each test, longer if you can. If Prime 95 generates errors, if the PC BSODs or crashes, or if the CPU temp exceeds 98°C (Tmax for your CPU is 100°C) then stop the test and let us know what happened.

A stable and properly cooled CPU and RAM should be able to run Prime95 pretty much indefinitely.

 

[luckzada]

So i tried running Prime95 with stock settings and it almost instantly crashes and if i try to run again the PC BSOD with error "MEMORY_MANAGEMENT".

When i run with stock settings the PC doesn't take long to BSOD, even without doing any stress test.

Link to the dump: https://files.catbox.moe/20odm4.dmp

 

[ubuysa]

That's almost certainly BAD RAM I think. Try removing on stick and run on just the other for long enough to have normally had a BSOD. Then swap the sticks over and run it long enough to have normally had a BSOD again. If it BSODs on both sticks on their own then it's not the RAM itself.

In that case I would run the Intel Processor Diagnostics Tool and see what that says about your CPU.

 

[luckzada]

I tried this last week and although it didn't BSOD, it crashed Cinebech instantly or in a few runs, which is a common symptom on when my PC is unstable. I tried with each RAM and i tried swapping slots as well but it didn't work out.

I ran Intel Diagnostic Tool yesterday with default settings on BIOS (100 BCLK) and it didn't return any errors.

I have some questions, it will be of great help if you can answer:

1. RAM is not on QVL list from the board, does this might explain the fact that it doesn't work even on stock settings or is probably faulty?

2. This issue started after a BIOS update and, according to the manufacturer website, the update cannot be rollback. Could this be a compatibility error, between the board and RAM?

3. On HWiNFO64 each RAM stick is running on different voltages, even if i enable XMP and the BIOS set to "Sync All PMICs", is this normal? Screenshot: https://files.catbox.moe/gcbh14.jpeg

 

[ubuysa]

1. RAM is not on QVL list from the board, does this might explain the fact that it doesn't work even on stock settings or is probably faulty?

The QVL lists RAM that has been tested and verified as compatible. RAM not on the QVL may well work, but on the other hand it may not. It's always wise to stick to QVL RAM to avoid this kind of is it compatible or isn't it compatible situation. We can't know whether your RAM is incompatible or not.

2. This issue started after a BIOS update and, according to the manufacturer website, the update cannot be rollback. Could this be a compatibility error, between the board and RAM?

You'd need to ask the manufacturer whether the BIOS update could be the problem. It's always wise to ask the board vendor whether a given BIOS update is suitable for your build before flashing it. My gut feel however, is that this is probably not the cause of your problems.

3. On HWiNFO64 each RAM stick is running on different voltages, even if i enable XMP and the BIOS set to "Sync All PMICs", is this normal? Screenshot: https://files.catbox.moe/gcbh14.jpeg

Is your RAM in matched pairs, or are they from different vendors or have different part numbers? Mismatched RAM is a casue of countless problems and BSODs. You should always buy RAM as a pack of macthed sticks. If the voltages are different then these are not a matched pair of RAM sticks and that is most liklely your problem. The only soution there is to junk these two sticks and buy a pack of two matched RAM sticks.

 

[luckzada]

Is your RAM in matched pairs, or are they from different vendors or have different part numbers? Mismatched RAM is a casue of countless problems and BSODs. You should always buy RAM as a pack of macthed sticks. If the voltages are different then these are not a matched pair of RAM sticks and that is most liklely your problem. The only soution there is to junk these two sticks and buy a pack of two matched RAM sticks.

They were suppose to be matching cause i bought the kit 2x8 GB from Kingston (KF560C40BBAK2-16), but even on stock settings they still have differ in 0.1v-0.2v.

So i am gonna try to contact Kingston to see what we can do about it, it is still on warranty.

 

[luckzada]

So, i discovered some things:

1. If i keep my CPU power capped at like 200W, i can boot and run stress tests, without crashes and BSOD, but then the PC just turns off. If i disable turbo and keep the CPU at base clock, the stress test runs fine as well.

2. On BIOS my "VRM core voltage", at stock, is 1.465, i saw some videos that was like 1.38 and even 1.28 at stock settings and when i lower the BCLK the voltage displays at 1.16. Is the stock voltage acceptable?

3. On BIOS, if i switch the microcode to "0x104 Microcode" instead of "Current microcode", the system becomes way more stable at 100.0 BCLK, it can run some stress test even with RAM at 6000MHz, but still has errors on Prime95 and some crashes overall.

Can this be temperature problems or maybe some problems with the motherboard?

 

[ubuysa]

You need to run everything at stock frequencies and voltages when troubleshooting BSODs. That means no CPU overclock or Turbo, and no RAM overclock (via XMP or similar), remove any undervolting you may be doing, and set all other BIOS settings to their defaults. It's very important that the system is running at the basic design specs when troubleshooting BSODs. If it fails Prime95 or Memtest86 at those default settings then it's a hardware problem. If it's Prime95 failing then it's most likely the CPU.

 

[luckzada]

Ok, i will do some more tests, but in worst case scenario i will just keep at 96.0 BCLK, the perfomance hit is nothing significant.

But, thank you so much for answering everything, you really helped me a lot trying to discover what the problem is, i really appreciate it.