[SOLVED] PC can run benchmarks but loading games causes crashes ?

[CaribRinus]

Out of nowhere, my computer crashes to desktop when loading games. With repeated tries leading to black screens or a rare BSOD. The weird thing is, running Unigine's Superposition benchmarks both in Performance as in Game mode does not lead to crashes. Every other game I try to load, it either does not load or gets to the main menu but can't load maps. Before, I could run every game with reasonable settings without issues, which makes sense if you look at my specs.

When the games crash to desktop, usually they refer to an issue with the video memory or video device.

I can replicate the crash by using Prime95 to stress test the CPU.

Specs:

1. CPU: Intel Core i9-13900KF
2. GPU: MSI RTX 3070 Ti Ventus 3X OC
3. RAM: 2x Kingston FURY Beast 16 GB 5200 MHz DDR5
4. MoBo: GIGABYTE Z690 AORUS ELITE AX
5. PSU: Seasonic Focus GX-850
6. SSD: Samsung 980 Pro 2TB M.2 NvME
7. CPU Cooler: be quiet! Dark Rock Pro 4

What have I already done? (not in chronological order)

1. Reseated GPU
2. Gave GPU independent PCi-e cables (before it was a single bridged cable)
3. Ran memtest86 with 4 passes 0 errors
4. Reinstalled Windows 10, upgraded to 11, reinstalled (not clean) back to 10
5. Updated GPU Drivers
6. Reinstalled GPU Drivers (DDU)
7. Confirmed PSU Wattage is sufficient using NewEgg's PSU Calculator
8. The 2 BSOD errors were irql_not_less_or_equal.
9. Analyzed the only 2 BSOD minidumps
   1. First one mentioned something with Intel.sys
   2. Second one mentioned an Nvidia driver
10. Ran Intel's Processor Diagnostic Tool without issues
11. DISM Restorehealth and sfc /scannow
12. Monitoring temperatures during benchmarks (CPU Avg 90 C GPU Avg 65 C)
13. Ran SeaTools Short, 2 min generic and Long Self tests with 0 errors
14. Updated Bios to most recent version F27

What have I done in the past 6 months?:

1. Reapplied CPU thermal paste
2. Got a new CPU, because my previous one was faulty

I have nothing else to troubleshoot. I want to be sure I have tried everything I can to identify if it's a hardware issue - and if so, which part - so that I know what steps to take to fix this issue. I just can't imagine the CPU being faulty. That would mean two CPU's I have bought in short succession are faulty. The chances of that happening are slim as I've read in other threads.

Any help would be appreciated.

 

[helper800]

When you load a game you tax your storage devices as well as the CPU and RAM. You also start using the graphics card after the loading is completed. It is odd that you have 1 RAM stick, but this should not cause any issues, especially if it passes memtest. What slot from the CPU socket is your RAM stick slotted into? I would check your storage devices for weird performance or operation.

 

[rgd1101]

what version of windows?
latest motherboard bios?
cpu is bit warm

 

[CaribRinus]

When you load a game you tax your storage devices as well as the CPU and RAM. You also start using the graphics card after the loading is completed. It is odd that you have 1 RAM stick, but this should not cause any issues, especially if it passes memtest. What slot from the CPU socket is your RAM stick slotted into? I would check your storage devices for weird performance or operation.

Ah my mistake, updated post to reflect the two sticks of RAM.

I have ran both SeaTools 2 min generic as self long test with no errors. Updated post.

 

[CaribRinus]

what version of windows?
latest motherboard bios?
cpu is bit warm

Windows 10
Updated to latest Bios
I agree that the CPU is warm, but it is not hot when it crashes, which is usually the moment higher loads are put on it.

 

[helper800]

I agree that the CPU is warm, but it is not hot when it crashes, which is usually the moment higher loads are put on it.

Is there any way to check VRM temperatures? If not try limiting the CPU power limits in BIOS and test.

 

[scout_03]

see if hwinfo wil read it also made sure cpu cooler fans is running as it should .

 

[CaribRinus]

Is there any way to check VRM temperatures? If not try limiting the CPU power limits in BIOS and test.

VRM Mos barely hitting 40 C

see if hwinfo wil read it also made sure cpu cooler fans is running as it should .

All fans have reasonable RPM's. I also tried running Prime95 with the fans all set to Full Speed in Bios, but still the crash occurs. This time, I have gotten a BSOD.

Using WinDbg I can debug the following:

1. SYSTEM_SERVICE_EXCEPTION (3b)
   An exception happened while executing a system service routine.
2. Caused by: GenuineIntel.sys (IP_MISALIGNED_GenuineIntel.sys)

Second BSOD of the night is a IRQL_NOT_LESS_OR_EQUAL pointing to ntkrnlmp.exe

 

[helper800]

VRM Mos barely hitting 40 C


All fans have reasonable RPM's. I also tried running Prime95 with the fans all set to Full Speed in Bios, but still the crash occurs. This time, I have gotten a BSOD.

Using WinDbg I can debug the following:

1. SYSTEM_SERVICE_EXCEPTION (3b)
   An exception happened while executing a system service routine.
2. Caused by: GenuineIntel.sys (IP_MISALIGNED_GenuineIntel.sys)

Second BSOD of the night is a IRQL_NOT_LESS_OR_EQUAL pointing to ntkrnlmp.exe

Try a clean boot and see if you still have crashes.

 

[CaribRinus]

Try a clean boot and see if you still have crashes.

Doesn’t seem to crash when in safe mode while running Prime95. At least far out of the same time span as before (10 mins vs 30 seconds)

 

[helper800]

Doesn’t seem to crash when in safe mode while running Prime95. At least far out of the same time span as before (10 mins vs 30 seconds)

This is a small tutorial on how to do a clean boot. This is running Windows without any other services or processes than necessary.

 

[CaribRinus]

This is a small tutorial on how to do a clean boot. This is running Windows without any other services or processes than necessary.

Crash occurs with Prime95 with clean boot as with external drivers enabled in same time span. No BSOD, just a black screen.

I find it odd that it does work in safe mode.
I have a feeling that my clean install of windows was not performed correctly. I will be doing that next. Thanks for all the help so far.

Update:

I did a clean install, to no avail. CPU temps consistently hitting 100C+ during stress testing and crashing. Could be a temperature issue. I ordered a new liquid CPU cooler because I wanted to get one either way. Will keep this post updated.

Any other suggestions happily welcome

 

[helper800]

Crash occurs with Prime95 with clean boot as with external drivers enabled in same time span. No BSOD, just a black screen.

I find it odd that it does work in safe mode.
I have a feeling that my clean install of windows was not performed correctly. I will be doing that next. Thanks for all the help so far.

Update:

I did a clean install, to no avail. CPU temps consistently hitting 100C+ during stress testing and crashing. Could be a temperature issue. I ordered a new liquid CPU cooler because I wanted to get one either way. Will keep this post updated.

Any other suggestions happily welcome

In the mean time I would try reinstalling your air cooler. What could be happening is that certain parts of the CPU are being cooled properly, but because of a bad mount or not enough thermal grease other parts are not. This can happen because of uneven pressure during the mounting process.

 

[CaribRinus]

SOLVED:

It was indeed a temperature issue. Got a liquid CPU cooler, added 3 extra fans to my case and now it seems like no crashes are occuring. Thanks everyone for thinking with me.