Question Random BSODs ntoskrnl.exe, more frequent on startup and during heavy usage like gaming ?

[Recondite]

Hi,

I have been facing this issue since past 10 months on my Asus TUF A15 2021- FA506IC Laptop.

Specs:
Ryzen 7 4800H
RTX 3050
16GB RAM
WD SN570 500GB M.2 NVMe SSD

What I did:

• Clean installed Windows (three times)
• Updated all the drivers
• Ran Windows Memory Diagnostic and MemTest- Result: All good
• Stress tested CPU and GPU- Result: All good
• While testing SSD using Crystal disk mark BSOD(ntoskrnl.exe) was happening again and again. So I changed my SSD, after changing it Crystal disk mark ran successfully but main issue still didn't get solved
• Changed the laptop's power adapter (in case of power issue)

Link to recent 5 system crash minidump files:
minidump

 

[ubuysa]

The dumps all point very strongly at RAM. There is no indication of any of these BSODs being third-party driver related. I appreciate that you've tested your RAM, did you do so with Memtest86 (free)? If not then please do so. Be aware that no memory tester can ever prove that your RAM is good, only flag it as bad.

When you run Memtest86, if no errors have been flagged after the four iterations of the 13 different tests that the free version does, then restart Memtest86 and do another four iterations.

Even if Memtest86 finds no errors, and if you have more than one RAM stick, then remove on stick of RAM and run like that for a few days. Then swap sticks for a few days. This will clearly show whether you have a flaky stick.

• When you reinstalled Windows, from where did you source the drivers?
• When you updated drivers, how did you do that?
• Have you monitored the temperatures at all? If not, download HWMonitor (free) and post a screen shot of CPU and GPU temps both at idle and at full load. Laptops are notorious for overheating.
• How did you stress test the CPU and GPU?

 

[Recondite]

The dumps all point very strongly at RAM. There is no indication of any of these BSODs being third-party driver related. I appreciate that you've tested your RAM, did you do so with Memtest86 (free)? If not then please do so. Be aware that no memory tester can ever prove that your RAM is good, only flag it as bad.

When you run Memtest86, if no errors have been flagged after the four iterations of the 13 different tests that the free version does, then restart Memtest86 and do another four iterations.

Even if Memtest86 finds no errors, and if you have more than one RAM stick, then remove on stick of RAM and run like that for a few days. Then swap sticks for a few days. This will clearly show whether you have a flaky stick.

• When you reinstalled Windows, from where did you source the drivers?
• When you updated drivers, how did you do that?
• Have you monitored the temperatures at all? If not, download HWMonitor (free) and post a screen shot of CPU and GPU temps both at idle and at full load. Laptops are notorious for overheating.
• How did you stress test the CPU and GPU?

- Installed all the drivers from manufacturer's official website i.e Asus
- I just typically installed the files that I downloaded from the same
- Yes, I think temps are quite normal considering a gaming laptop
Here are the screenshots:
screenshots
- Used various softwares like, CinebenchR23, Heaven Benchmark, 3d mark time spy, Aida64 etc

Additionally, After Memtest86 I removed one stick of RAM that I had myself installed but the problem still persisted but I hadn't tested removing the stock RAM installed by the manufacturer. As I don't have any other RAM stick, I'll replace the manufacturer's RAM with mine and try using it on 8GB only, I'll update you about the results.

I have also used linux and faced issues like freezes/unresponsive black screens also

Here are some more minidump files of recent crashes:
moreminidump

The dumps all point very strongly at RAM. There is no indication of any of these BSODs being third-party driver related. I appreciate that you've tested your RAM, did you do so with Memtest86 (free)? If not then please do so. Be aware that no memory tester can ever prove that your RAM is good, only flag it as bad.

When you run Memtest86, if no errors have been flagged after the four iterations of the 13 different tests that the free version does, then restart Memtest86 and do another four iterations.

Even if Memtest86 finds no errors, and if you have more than one RAM stick, then remove on stick of RAM and run like that for a few days. Then swap sticks for a few days. This will clearly show whether you have a flaky stick.

• When you reinstalled Windows, from where did you source the drivers?
• When you updated drivers, how did you do that?
• Have you monitored the temperatures at all? If not, download HWMonitor (free) and post a screen shot of CPU and GPU temps both at idle and at full load. Laptops are notorious for overheating.
• How did you stress test the CPU and GPU?

 

[ubuysa]

If you also experience problems with Linux on this PC then it's absolutely definitely a hardware problem.

The temps do look fine, especially for a laptop.

RAM or CPU are the common hardware culprits but it really could be anything. Swapping RAM and Memtest86 are the best ways to test RAM. With an AMD CPU you're really only left with Prime95 to stress the CPU, but be careful with that on a laptop, it makes the CPU run VERY hot. If you've never run Prime95 before then ask.

Do Asus have any hardware diagnostic apps for your laptop?

 

[Recondite]

If you also experience problems with Linux on this PC then it's absolutely definitely a hardware problem.

The temps do look fine, especially for a laptop.

RAM or CPU are the common hardware culprits but it really could be anything. Swapping RAM and Memtest86 are the best ways to test RAM. With an AMD CPU you're really only left with Prime95 to stress the CPU, but be careful with that on a laptop, it makes the CPU run VERY hot. If you've never run Prime95 before then ask.

Do Asus have any hardware diagnostic apps for your laptop?

I never ran Prime95 before. Hardware diagnostic app says all good.
I was wondering how would I find if it's the motherboard who is culprit?

 

[ubuysa]

Before we test the CPU I'd like to look at the CPU cooling, especially since these problems seem to happen when it's under load. Download HWMonitor (free) and post two screenshots of the CPU temps (expand this section to show each core). Post one at idle and the other at the maximum load you can throw at it.

What make, model, and size of PSU do you have?

 

[Recondite]

Before we test the CPU I'd like to look at the CPU cooling, especially since these problems seem to happen when it's under load. Download HWMonitor (free) and post two screenshots of the CPU temps (expand this section to show each core). Post one at idle and the other at the maximum load you can throw at it.

What make, model, and size of PSU do you have?

As it's a laptop. It has a 180W adaptor that comes inside the box with laptop.
Here are the required screenshots:
ss
Apologies for the late reply.

 

[ubuysa]

I'd forgotten that it was a laptop, but those temps are fine.

To test the CPU I suggest you run Prime95, however this will make you're CPU run very hot so you definitely want to be running a temperature monitor (like HWMonitor) whilst Prime95 is running. Also be sure the laptop vents and heat exchanger (next to the fan(s) are completely clean, raise the laptop to allow plenty of air in underneath it, place it on a metal surface rather than wood or plastic (metal doesn't retain heat like wood and plastic do), and ensure that there is plenty of room around the laptop for cool air to flow in and hot air to escape.

Run all three Prime95 tests (small FFTs, large FFTs, and Blend) for an hour each test - if you can. If Prime95 generates errors, if the system crashes or BSODs, or if the CPU temp exceeds 100° C (Tmax for your Ryzen 7 4800H is 105° C) then stop Prime95 and let us know what happened.

 

[Recondite]

Here's what happened during these tests:
- Ran smallest FFTs for 1 hour and so. CPU temps were quite fine i.e. below 97° C. Got no BSOD.
- Ran small FFTs again for 1 hour. CPU temps were same. Got no BSOD.
- Ran large FFTs and there were the same BSODs that I usually get. As it's a RAM heavy test. I thought I should run a customized test that's more heavy on RAM. Found a relevant Reddit thread which suggested these settings:
- unchecked in-place
- range from 448K to 409
- allocated 70% of my RAM
same BSOD happened on this test too.
- Haven't ran Blend after this

 

[ubuysa]

We already pretty much knew that there is a hardware problem on here, this just confirms that. You're right that the large FFT tests stresses RAM more than the small FFT tests.

Earlier you wrote this...

Additionally, After Memtest86 I removed one stick of RAM that I had myself installed but the problem still persisted but I hadn't tested removing the stock RAM installed by the manufacturer. As I don't have any other RAM stick, I'll replace the manufacturer's RAM with mine and try using it on 8GB only, I'll update you about the results.

It's important that RAM is installed in matched pairs, so I would suggest replacing both RAM cards with a matched pair in a single pack - do ensure that they are compatible with your laptop model before you buy them. I'm not guaranteeing that this will solve your problem but the alternative is replacing the CPU or the motherboard - for a laptop that really means a new laptop.