Question Computer "crashes" under load ?

[Ok-Somewhere-9161]

Hello and happy New Year everyone.


I've been fighting an issue with my desktop for some time and have finally decided to reach out for help. About a year or so ago my computer began to "crash" when playing some games, generally newer ones. One such game is Resident Evil 2 Remake, a game that my PC was able to play before without issue but then began to experience this issue.

I put "crash" in quotation marks as it's not a full crash where the entire system powers down. What happens is that my keyboard, mouse, and monitors all power down but the tower itself seems to still be running (there are no beeps from the motherboard when this happens). The fans are going and the lights on the motherboard are going as per normal; however, I did notice the debug light for the CPU was a solid red color.

The only way to get out of this state is to press the power button and do a hard shutdown. Unplugging and plugging the monitors and keyboard/mouse don't do anything. This is a custom-built PC that did well for awhile but randomly started up this issue one day. It proceeded to get worse, to the point that having Twitch or two tabs with YouTube open could cause the "crash" to happen. The steps I've taken to remedy this have led to some minor improvement but it is still happening. Most recently, I tried running a full system scan with my anti-virus and the "crash" happened again. In this way it seems to be repeatable and not with a graphically intensive item.

The system does boot up perfectly fine, I'm able to run some games on it and run other anti-malware and anti-virus scans without issue.


Here are the specs for the computer:

Motherboard: MSI MAG X570S Tomahawk Max Wifi

CPU: AMD Ryzen 5 2600X Six-Core Processor 4.05 GHz

RAM: 16 GB DDR4 (G-Skill Ripjaw 3200 2x8 GB [model: F4-3200C16-8GVKB])

OS: Windows 11 Home

2 SSDs (250GB and 1 TB)

GPU: GeForce RTX 2060

PSU: Corsair CX750M (750 Watt)

2 monitors using Display Port

Never done overclocking


With all that said, here are the things I've attempted to try and fix this problem since it started. Most of these either had no discernible effect or led to minor improvement.

Updated OS from Win 10 to Win 11

Updated OS and all drivers, including BIOS

Replaced the motherboard

Replaced the GPU which was under warranty

Replaced the PSU. Used both NewEgg and Seasonic's calculators to verify that 650W was what they suggested, I put a 750W in there to be safe.

Ensured PC is in well-ventilated area, sprayed entire inside with condensed air, removed the side panel of the tower to ensure air circulation

Replaced thermal paste on CPU

Updated power settings to favor performance and limit CPU usage to 95% max

Used Core Temp and NVidia overlay to confirm temperatures on CPU and GPU are not excessive (crash tends to happen when both are still in the 60-70 C range)

Booted in Safe Mode

Temporarily turned off all anti-virus, VPN, and firewall protection

Turned off services that aren't necessary

Checked Reliability Monitor and Event Viewer for issues only to find nothing except for "Windows did not properly shut down" from the hard shut down I do manually

Run MemText86 for several hours to verify the RAM is good

Verified the RAM is compatible with the motherboard

Run CHKDSK on both SSDs to verify no issues with either drive

Run UserBenchmark without issue. Lowest item was the CPU at 89.2%

Downloaded HWInfo, I can provide info from that if needed.

There may be some other steps I'm forgetting since it's been an uphill battle for a while. Any suggestions would be helpful and I'm open to providing more information if it helps solve the problem. Even an explanation of what this "crash" is that doesn't fully shut down the computer would be appreciated.

 

[Ok-Somewhere-9161]

Small update: I should note that I have also gone through the NO POST checklist. Tomorrow I'm going to try moving my computer to a different room to test it with a different outlet, no power strip, and only one monitor to see if that works.

Also worth noting is that my motherboard does seem to have a CPU PWR 2 port that's a 4-pin connection (next to the 8-pin CPU PWR 1) that is currently not plugged in. Looking into it seems to indicate that this port shouldn't be needed unless I was doing really heavy loads like overclocking and the 8-pin I have should be able to handle most everything. I've ordered an extension to turn the 8-pin cable into an 8 and 4-pin.

If neither of those work then I'm really not sure what else I can do short of buying a new CPU.

 

[kira-faye]

If you've done all that it's basically down to the CPU itself or the RAM (memtest86 results notwithstanding). Borrow some if you can, order a kit from somewhere with a good return policy if you can't.