Mysterious crashes on a computer whose components all seem fine

[Alexko]

Hi everyone!

My brother's laptop crashes almost systematically in most games, after a few minutes of gameplay. The weird thing is that I can't identify any software or hardware issue that could explain this.

Why I think the software's fine: He did a full reinstall of Windows, with fresh, up-to-date drivers. Games still crash.

Why I think the GPU's fine: Thinking the GPU might be overheating or faulty somehow, I had him run Furmark, which ran fine. I also had him try one of his games on the iGPU instead of the dGPU, and it crashed just the same.

Why I think the CPU's fine: I had him run OCCT as well, and it ran without issues. I also had him run both Furmark and OCCT at the same time, which caused no errors, nor any overheating, meaning the power supply and cooling systems are probably all fine too.

Why I think the RAM is fine: I had him run memtest86 over the full 48 tests (about 3 hours): no errors.

Why I think the SSD is fine: I had him install a game on a flash drive and run it from there: it crashed just the same. But the OS and drivers are still running from the SSD, so maybe that's the culprit?

So I'm down to either the motherboard or the SSD, maybe through some issue that somehow pops up when the graphics driver is doing its thing (sometimes, not with Furmark) and possibly accessing stuff from the drive, but somehow not in other kinds of applications, since he's not really seeing any crashes outside of games. Then again, he doesn't do anything too demanding with his computer aside from gaming.

If the motherboard is faulty, then it's a fault that does not get triggered by intensive CPU usage, or RAM usage. I guess maybe something wrong with the PCIe bus? But then it probably wouldn't crash when gaming on the iGPU.

Does anyone have any bright ideas? If the motherboard is at fault, how could I confirm it? Or am I missing something about another component? I'm not sure if that actually adds any valuable information, but 3D Mark tends to crash just like games, and unlike Furmark.

The computer is an Acer laptop, specifically a Helios 300, with an i7-9750H, RTX 2060, and 16GB of DDR4-2666 (SK Hynix chips, apparently). Naturally, the warranty has expired.

 

[pharma]

The "Event Viewer" doesn't give any information about the crashes?

Edit: Does it "crash" out of the game only or does it force an OS reboot?

Sometimes heat is a problem with older laptops where dust buildup inside exhaust vents can cause additional stress.

 

[digitalwanderer]

When you say the CPU/GPU "ran fine", what kind of temperatures are we talking? Also did you take off all the bloatware, what applications are starting on start up?

Oh, and PSU could be the culprit too. If it ain't up to the constant full draw it could manifest like a crash.

 

[Alexko]

The "Event Viewer" doesn't give any information about the crashes?

Warnings like "Loading driver\WudfRd failed for peripheral X" where X is the GPU, so in the case of the latest crash, the Intel iGPU.

And errors like "the computer restarted after error checking"

And "Dump file creation failed due to an error during dump creation", because I guess a dump file might be too useful.

 

[Alexko]

When you say the CPU/GPU "ran fine", what kind of temperatures are we talking? Also did you take off all the bloatware, what applications are starting on start up?

Oh, and PSU could be the culprit too. If it ain't up to the constant full draw it could manifest like a crash.

We're talking about high temperatures during stress tests, around 80 to 85C, which is still within specs, I believe. During games, things are cooler.

Running Furmark and OCCT at the same time to max out power consumption did not cause any crashes, which is why I figured the PSU was probably fine, since games aren't as demanding as that. And apparently Crusader Kings 3 crashes a lot and usually fast, even though it's not exactly Crysis.

My brother reinstalled Windows, so there shouldn't be any bloatware. I'll ask about the apps running on start-up.

 

[iroboto]

Can you update the firmware on the mobo? After reading your tests, possible issue with data transfer between system memory and the gpu vram? not sure why igpu is crashing though.

Can you also try changing your display port? seeing if that makes any difference?

last thing I can think of, the swap file/virtual memory failure.

cpu could still be an issue. laptops may detect it's getting burned in for tests and lower clock speeds to ensure thermals stay put. This may not catch for games.

Maybe worth while to detune the CPU as much as possible and see if it still crashes during games?

 

[Malo]

Make sure all BIOS and system drivers are latest versions, especially chipset drivers?

 

[digitalwanderer]

Can you also try changing your display port? seeing if that makes any difference?

I can answer that one; no, it's a laptop.

 

[iroboto]

I can answer that one; no, it's a laptop.

ahh... they don't have 2 ports? 1 HDMI and 1 Display port?

 

[WhiningKhan]

I have a desktop PC with old HD7870XT which has overheated for long periods of time due to fan failure. It runs some games just fine(Elite Dangerous original and Horizons, many old games), some games only a couple of minutes (Troy:Total War, ED Odyssey) and some games I was able to stabilize with forced frame rate limit (Civ6).

It is in daily use for non-gaming purposes with no problems.

 

[digitalwanderer]

Do they change input devices for gaming? Or do they normally use a mouse with their laptop like a sane person?

 

[digitalwanderer]

Also I'm doubting it's the mobo if it can run all those benchmarks fine.

 

[iroboto]

you're thinking to force frame rate limiters to ensure the issue doesn't crop up?

 

[Alexko]

Can you update the firmware on the mobo? After reading your tests, possible issue with data transfer between system memory and the gpu vram? not sure why igpu is crashing though.

Can you also try changing your display port? seeing if that makes any difference?

last thing I can think of, the swap file/virtual memory failure.

cpu could still be an issue. laptops may detect it's getting burned in for tests and lower clock speeds to ensure thermals stay put. This may not catch for games.

Maybe worth while to detune the CPU as much as possible and see if it still crashes during games?

Chipset driver updated, to no avail. I guess the BIOS would be worth a try. I can't see much of anything else to update. It's a laptop with internal monitor, so no display output is involved. What do you have in mind regarding CPU detuning? Just underclocking it?

Make sure all BIOS and system drivers are latest versions, especially chipset drivers?

Done for drivers, will do for BIOS.

Do they change input devices for gaming? Or do they normally use a mouse with their laptop like a sane person?

Same mouse as ever, and the integrated keyboard, as usual. No change whatsoever.

you're thinking to force frame rate limiters to ensure the issue doesn't crop up?

Worth a try, I guess!

 

[Silent_Buddha]

Hi everyone!

My brother's laptop crashes almost systematically in most games, after a few minutes of gameplay. The weird thing is that I can't identify any software or hardware issue that could explain this.

Why I think the software's fine: He did a full reinstall of Windows, with fresh, up-to-date drivers. Games still crash.

Why I think the GPU's fine: Thinking the GPU might be overheating or faulty somehow, I had him run Furmark, which ran fine. I also had him try one of his games on the iGPU instead of the dGPU, and it crashed just the same.

Why I think the CPU's fine: I had him run OCCT as well, and it ran without issues. I also had him run both Furmark and OCCT at the same time, which caused no errors, nor any overheating, meaning the power supply and cooling systems are probably all fine too.

Why I think the RAM is fine: I had him run memtest86 over the full 48 tests (about 3 hours): no errors.

Why I think the SSD is fine: I had him install a game on a flash drive and run it from there: it crashed just the same. But the OS and drivers are still running from the SSD, so maybe that's the culprit?

So I'm down to either the motherboard or the SSD, maybe through some issue that somehow pops up when the graphics driver is doing its thing (sometimes, not with Furmark) and possibly accessing stuff from the drive, but somehow not in other kinds of applications, since he's not really seeing any crashes outside of games. Then again, he doesn't do anything too demanding with his computer aside from gaming.

If the motherboard is faulty, then it's a fault that does not get triggered by intensive CPU usage, or RAM usage. I guess maybe something wrong with the PCIe bus? But then it probably wouldn't crash when gaming on the iGPU.

Does anyone have any bright ideas? If the motherboard is at fault, how could I confirm it? Or am I missing something about another component? I'm not sure if that actually adds any valuable information, but 3D Mark tends to crash just like games, and unlike Furmark.

The computer is an Acer laptop, specifically a Helios 300, with an i7-9750H, RTX 2060, and 16GB of DDR4-2666 (SK Hynix chips, apparently). Naturally, the warranty has expired.

If you have GeForce Experience on that machine try disabling or uninstalling it. I've had a few people who resolved crashes during their games by uninstalling GeForce Experience.

Regards,
SB

 

[orangpelupa]

So the error on event viewer says it's the Intel gpu?

Try manually updating the Intel gpu driver by manually downloading the zip from Intel download center and manually use device manager to install it.

Also try to dig into the uefi and disable the Intel gpu

 

[Alexko]

So the error on event viewer says it's the Intel gpu?

Try manually updating the Intel gpu driver by manually downloading the zip from Intel download center and manually use device manager to install it.

Also try to dig into the uefi and disable the Intel gpu

It said so in this case because my brother tried running the game on the iGPU instead of the dGPU, to see if the dGPU was responsible for the crashes.

If you have GeForce Experience on that machine try disabling or uninstalling it. I've had a few people who resolved crashes during their games by uninstalling GeForce Experience.

Regards,
SB

Worth a try, but since crashes also happen on the iGPU, I'm not too optimistic about that one.

Also, he tried underclocking the GPU, but it didn't help, the game still crashed. It seems his BIOS may be completely locking the CPU clock, however, because there's no option in there to change it, and it's grayed out in ThrottleStop. He can however disable Turbo, so he'll try that next.

I should perhaps mention something super weird: before this current situation where games pretty much always crash, things used to be stable, then from February/March onwards, my brother noticed that he would get crashes whenever a new update was available, but not yet installed for his graphics driver. After updating the driver, however, the crashes would disappear, so he didn't worry too much about it. Then, with a new update in September, the crashes just didn't go away.

This is super strange so he just uninstalled the driver entirely and reinstalled it from scratch, to no avail. When he mentioned this to me, I thought this sounded like a weird, but deep software issue, so I recommended that he reinstall Windows, which he did, but also in vain. I honestly have no idea what to make of this. I mean, I suppose that the availability of a new update could possibly trigger bugs in the driver, but even after reinstalling Windows, and even when running games from the iGPU?

 

[orangpelupa]

but even after reinstalling Windows, and even when running games from the iGPU?

AFAIK nvidia + intel laptops basically married both GPU into one (in certain aspects). so maybe even if the dGPU that's borked, the error will reports as the iGPU that's borked.

btw have you tried using manual install of intel GPU driver? the intel download center here https://www.intel.com/content/www/us/en/download-center/home.html you would need to grab the zip instead of the exe driver.

anyway, it also could be components issues on the motherboard so need to check each components one by one or replace motherboard. My lenovo laptop even got its motherboard replaced because broken fingerprint reader (i was like.. huh? what?)

 

[digitalwanderer]

Try DDU.

 

[Davros]

btw have you tried using manual install of intel GPU driver?

Id suggest removing the driver and disabling onboard graphics in the bios - would be the better option