Question Inconsistent crashes when PC is under load ?

[Jacob Harris12]

My computer was running perfectly fine until a few days ago when it started crashing shortly after getting into high load games. For example in Forza Horizon 5 sometimes it crashes right when I load into the game and sometimes it can take a min. Sometimes I can play however long I want and it doesn't crash. I have tried testing each individual stick of RAM and it crashed on each one. I did a fresh install of windows and updated any drivers I could think of. Updated my bios which also did nothing. The last thing I tried was swapping in a 1000w PSU but the crashing continues.

The whole time I have had this issue I also haven't been able to get any crash dumps. I set up my system to create crash dumps and have WhoCrashed to read them but none of the crashes created a dump.

My system specs...

CPU: Ryzen 3800xt

GPU: Radeon 6950xt

RAM: 4x8GB Corsair Vengeance Pro

MB: Asus B550F Gaming

PSUs: Aresgame 850w 80+ Gold and the one I swapped in was a Thermaltake Toughpower GF3 1000w 80+ Gold

I'm about out of ideas so any help would be greatly appreciated.

 

[boju]

Not sure about dump files, searching a little with phrase crash dump file not created gives a few ideas why, might want to take a look.

What storage drive are you using? Check ssd with this guide.

How to Check SSD Health in Windows 10 and Windows 11

Find out how many writes your SSD has left.

www.tomshardware.com

 

[Misgar]

Is your 4x8GB from a matched kit of 4 DIMMs purchased as one item, or two pairs of similar/dissimilar DIMMs which have not been carefully matched by the manufacturer?

You don't mention the exact part number(s) for the memory, but if you're trying to run 4 DIMMs at their maximum rated XMP overclock speed, you may have problems.

It's usually much easier to overclock 2 DIMMs, not 4. Fitting 4 DIMMs places additional load on the Integrated Memory Controller bus and can reduce the maximum overclock, necessitating manual intervention and relaxed timings.

As a first step, if you're running XMP memory overclock, disable it and allow the BIOS to drop back to the default (lower) JEDEC speed, which is normally around 2133MT/s for DDR4.

Similarly, if you're overclocking the 3800XT, disable PBO, etc.

If you're boosting the GPU performance with MSI Afterburner, switch off the graphics card overclock.

The idea is to reduce the platform to a (potentially) more stable condition, then run the games/programs that give trouble.

If you have problems with no overclocking, remove 3 of the 4 DIMMs, create a bootable USB memory stick with MemTest86+ and test the first memory stick. Run a full cycle (approx 1 hour).

Repeat the RAM test until you have individually tested each one of the 4 DIMMs in isolation. If you encounter ANY errors at the JEDEC speed 2133MT/s, treat that DIMM as bad and discard/RMA.

If all 4 DIMMs pass MemTest86+ individually at JEDEC 2133MT/s (or possibly 2400MT/s) you can try testing them again with XMP enabled using MemTest86+. As before, run at least one full cycle of the test. This can take several hours if all 4 DIMMs are checked at the same time.

The 3800XT is specified for use with memory up to DDR4-3200, but you may encounter problems with 4 DIMMs at 3200MT/s or higher speeds. If you pull 2 DIMMs leaving just 2 DIMMs, 3200MT/s should be achievable, but is not guaranteed.

 

[Jacob Harris12]

Not sure about dump files, searching a little with phrase crash dump file not created gives a few ideas why, might want to take a look.

What storage drive are you using? Check ssd with this guide.

How to Check SSD Health in Windows 10 and Windows 11

Find out how many writes your SSD has left.

www.tomshardware.com

I have windows on a WD_BLACK SN750 NVMe SSD 1tb. I ran the SMART Diagnostic and it came back with no errors. And the drive shows 99% health.

 

[Jacob Harris12]

Is your 4x8GB from a matched kit of 4 DIMMs purchased as one item, or two pairs of similar/dissimilar DIMMs which have not been carefully matched by the manufacturer?

You don't mention the exact part number(s) for the memory, but if you're trying to run 4 DIMMs at their maximum rated XMP overclock speed, you may have problems.

It's usually much easier to overclock 2 DIMMs, not 4. Fitting 4 DIMMs places additional load on the Integrated Memory Controller bus and can reduce the maximum overclock, necessitating manual intervention and relaxed timings.

As a first step, if you're running XMP memory overclock, disable it and allow the BIOS to drop back to the default (lower) JEDEC speed, which is normally around 2133MT/s for DDR4.

Similarly, if you're overclocking the 3800XT, disable PBO, etc.

If you're boosting the GPU performance with MSI Afterburner, switch off the graphics card overclock.

The idea is to reduce the platform to a (potentially) more stable condition, then run the games/programs that give trouble.

If you have problems with no overclocking, remove 3 of the 4 DIMMs, create a bootable USB memory stick with MemTest86+ and test the first memory stick. Run a full cycle (approx 1 hour).

Repeat the RAM test until you have individually tested each one of the 4 DIMMs in isolation. If you encounter ANY errors at the JEDEC speed 2133MT/s, treat that DIMM as bad and discard/RMA.

If all 4 DIMMs pass MemTest86+ individually at JEDEC 2133MT/s (or possibly 2400MT/s) you can try testing them again with XMP enabled using MemTest86+. As before, run at least one full cycle of the test. This can take several hours if all 4 DIMMs are checked at the same time.

The 3800XT is specified for use with memory up to DDR4-3200, but you may encounter problems with 4 DIMMs at 3200MT/s or higher speeds. If you pull 2 DIMMs leaving just 2 DIMMs, 3200MT/s should be achievable, but is not guaranteed.

I have two kits of this https://www.amazon.com/gp/product/B07TC4TPCN/ref=ppx_yo_dt_b_asin_title_o07_s00?ie=UTF8&psc=1.

I know I shouldn't have two seperate kits in there so I pulled one of the kits out and ran them at stock speeds and still had the crashing. I am going to start the process of testing each stick now and will let you know what the results are.

 

[Jacob Harris12]

Is your 4x8GB from a matched kit of 4 DIMMs purchased as one item, or two pairs of similar/dissimilar DIMMs which have not been carefully matched by the manufacturer?

You don't mention the exact part number(s) for the memory, but if you're trying to run 4 DIMMs at their maximum rated XMP overclock speed, you may have problems.

It's usually much easier to overclock 2 DIMMs, not 4. Fitting 4 DIMMs places additional load on the Integrated Memory Controller bus and can reduce the maximum overclock, necessitating manual intervention and relaxed timings.

As a first step, if you're running XMP memory overclock, disable it and allow the BIOS to drop back to the default (lower) JEDEC speed, which is normally around 2133MT/s for DDR4.

Similarly, if you're overclocking the 3800XT, disable PBO, etc.

If you're boosting the GPU performance with MSI Afterburner, switch off the graphics card overclock.

The idea is to reduce the platform to a (potentially) more stable condition, then run the games/programs that give trouble.

If you have problems with no overclocking, remove 3 of the 4 DIMMs, create a bootable USB memory stick with MemTest86+ and test the first memory stick. Run a full cycle (approx 1 hour).

Repeat the RAM test until you have individually tested each one of the 4 DIMMs in isolation. If you encounter ANY errors at the JEDEC speed 2133MT/s, treat that DIMM as bad and discard/RMA.

If all 4 DIMMs pass MemTest86+ individually at JEDEC 2133MT/s (or possibly 2400MT/s) you can try testing them again with XMP enabled using MemTest86+. As before, run at least one full cycle of the test. This can take several hours if all 4 DIMMs are checked at the same time.

The 3800XT is specified for use with memory up to DDR4-3200, but you may encounter problems with 4 DIMMs at 3200MT/s or higher speeds. If you pull 2 DIMMs leaving just 2 DIMMs, 3200MT/s should be achievable, but is not guaranteed.

After testing each stick of ram passed Memtest86+

 

[boju]

Are you running multiple monitors? If you are, is there a difference in Hz? Wondering if graphics card might be unstable with the two connected. There was an issue with one of the recent Amd drivers, not sure if fixed yet, where multi monitor can cause instability. I think setting power plan in Windows to full performance sort of helped. Adrenaline doesn't have a setting to favor high performance does it? That's done with Windows power plan?

 

[Jacob Harris12]

Are you running multiple monitors? If you are, is there a difference in Hz? Wondering if graphics card might be unstable with the two connected. There was an issue with one of the recent Amd drivers, not sure if fixed yet, where multi monitor can cause instability. I think setting power plan in Windows to full performance sort of helped. Adrenaline doesn't have a setting to favor high performance does it? That's done with Windows power plan?

I only have one monitor. And yes in windows power plan you can change it to high performance. Also the crashing only occurs in some of my games. I have been playing Modern Warefare 2 for multiple hours with no problems.

 

[boju]

Have you tried high performance in these games that do crash?

Im suspecting graphics card driver atm, these games are struggling with, might have something to do with power draw and clock speeds causing card to be wonky, high performance might help steady it. Perhaps try older driver if you can if high performance doesn't help.

 

[Jacob Harris12]

Have you tried high performance in these games that do crash?

Im suspecting graphics card driver atm, these games are struggling with, might have something to do with power draw and clock speeds causing card to be wonky, high performance might help steady it. Perhaps try older driver if you can if high performance doesn't help.

I tried setting them to high performance and I went back and tried a hand full of drivers. Just for fun I upped the power draw limit on my card and nothing made a difference. I'm thinking about having a local repair shop take a look at it cause I'm running out of ideas.

 

[slayer_rohan]

I am getting PC crash after hardware upgrade as well

I recently added the Mobo+Cpu+Ram combo and my system has been crashing ever since.
CPU and GPU temps are normal. Can rule out overheating.

I have tried plugging it in different sockets in the house, crashed in multiple sockets

Event viewer critical error -
Event Id - 41, Task 63


System configuration:
CPU - i5-13600kf
32GB Ram - GSKILL Ripjaws V 32GB DDR4 3600 MHz CL18-22-22-42 1.35V
Motherboard - Z690 PG Riptide(latest Bios)
GPU - Gigabyte RTX 3080Ti 12G
OS - Windows 10 pro 64 bit (Build 19045)
SSD - Crucial MX500 500GB 2.5-inch SSD
HDD - WD Blue 1 TB 7200 RPM SATA 6 Gb/s 64 MB Cache
PSU - Cooler master v1000 Gold