[SOLVED] System freezing GPU -1% usage

[anaccountoutofdesperation]

So basically my PC is going wombo combo on freezes lately and with the specs and the age of this thing I didn’t expect it (not even 10 months old). There are two things that stand out about the freezes:

1. Never when doing high demand gaming, including playing Doom Eternal, Resi Village, Dark Souls 3, Fallout with mods, Cyberpunk 2077 etc.
2. Always when either nothing is happening or when I am on Word or Spotify or Chrome (reinstalled all including my RGB software such as Logitech G Hub, all Gigabyte applications including SIV, Easy Tune, and RGB Fusion). I then ran the PC without reinstalling this software and it still froze after some time). Freezes usually occur like once or twice a week and with work and study it is very immobilising .

I will go through a list of things I’ve done to solve this and a list of errors or things I’ve noticed with an untrained eye (new to this).

I got Malwarebytes…nothing found.
Diver Easy…everything is updated.
Uninstalled and reinstalled GPU software (Display adapters > NVIDIA GeForce RTX 3070 > Uninstall)

Reset BIOS to default and checked all Event Viewer logs only to see things which according to other forums are nothing, including:
“Audit events have been dropped by the transport. 0.”
Session "Cloud Files Diagnostic Event Listener" failed to start with the following error: 0xC0000022
“The application-specific permission settings do not grant Local Activation permission for the COM Server application with CLSID”

Ran “mdsched.exe” …nothing found.

Ran Cinebench…no overheating and full performance with no throttling from what I can tell.

Ran 3DMark…nothing, GPU at around 79 degrees Celsius full usage and idle at a cosy 45 degrees.

When one of the freezes happened I managed to get Task Manager up and my GPU was at 100% usage (this might be a false reading) but couldn’t load HWMonitor. Another time the GPU went to a strange -1% and then spiked to 100% but fans were practically idle and all other things including HDD, SSD, and CPU all around 51 Degrees with CPU usage around 10% and Ram and Storage at 0%.

Ran CrystalDiskInfo: health of both drivers in blue, listed as “good”.

I should reinstall windows but for the life of me I don’t want to unless I know it isn’t hardware related because backing up everything will require another hard drive. I can’t shake the feeling it is hardware related. GPU and PSU are latest things installed but the GPU was installed a month into purchasing the PC and never had a hiccup (8 months ago). I replaced an inadequate white grade 550W PSU with a 700W KOLINK ENCLAVE 700 W 80+ Gold and I installed the Cooler Master MasterLiquid ML240R so it may have started after this, but surely I would have noticed it sooner. The main reason I have posted this is because I know I could be missing some incompatibility or obvious sign that it is hardware related. If I have to reinstall Windows I will. I’m not avoiding any heavy lifting, but I’ve seen before how easily a bad MOBO, or other hardware issue can be identified.

PC Specs:

https://au.pcpartpicker.com/list/CLjRcT

The few parts not listed in the link:
Cooler Master MasterFan MF120 Halo ARGB 120mm
TP-Link TL-WN881ST 300Mbps Wireless N PCI Express Adapter
SilverStone Black 100mm PWM Fan Splitter Cable

 

[tennis2]

Presumed culprit = faulty stick of RAM.

Assuming you've got 2 sticks, you can test with 1 installed at a time.

 

[anaccountoutofdesperation]

Presumed culprit = faulty stick of RAM.

Assuming you've got 2 sticks, you can test with 1 installed at a time.

I made a mistake listing it as 2 x 8gb DDR4-3600 CL16 Memory. It's actually 1 x 16 GB DDR4-3000 CL16 Memory but irrespective that is easily replacable. I just thought if I ran Windows Memory Diagnostic it would have identified it as a RAM issue.

Cheers btw I will have to order some RAM and test this.

 

[tennis2]

I've had a known-faulty stick of RAM run for 19 hours on Prime95 before an error showed (that was with the other stick installed also, hence only testing 1 at a time). Folding@Home sussed it out in 10 minutes (that's actually how I was alerted to the issue)

 

[anaccountoutofdesperation]

I've had a known-faulty stick of RAM run for 19 hours on Prime95 before an error showed (that was with the other stick installed also, hence only testing 1 at a time). Folding@Home sussed it out in 10 minutes (that's actually how I was alerted to the issue)

Legend. I live in Aus and my state is only just coming out of lockdown so new RAM will have to be bought online but keen to give it a swap out now.
Thanks again.