Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Yeep
Nov 8, 2004
Problem description: A couple of months ago I bought Diablo IV and it would consistently cause my PC to blackscreen and reboot with a WHEA error 18 Cache Hierarchy error (with varying core numbers) after between 0 and 15 minutes of play. I had run the beta fine and this was the first time I'd seen this error. I had a spare PC where everything ran fine so figured it was just bad Blizzard code (other people on the Battle.net forums seemed to be having similar issues) so after trying a couple of fixes (BIOS/driver updates, frame rate limiting, removing the high-res texture pack) I gave up and used my spare PC for Diablo. Unfortunately that PC is needed elsewhere now and I've started getting the reboot in other games, notably Zwift which shouldn't be particularly taxing at all so I'd like to find the root cause.

Attempted fixes:
Updated Windows, GPU drivers, Motherboard drivers, BIOS, GPU firmware.
Unplugged 2nd monitor and all USB devices except KB/mouse.
DDU clean GPU driver install.
Ryzen curve optimizer (this made things worse).
Full BIOS reset. Disabling XMP profiles.
Re-seated GPU (both ends of riser cable), RAM and all PSU cables.
Blown dust out of everything.
memtest ran fine for a full run

I found some Reddit posts saying that sometimes 5900x just go bad and I should RMA it while it's still under warranty but AMD asked me to provide evidence of more troubleshooting and I'm not sure what else I can do short of re-installing Windows.

Switching the 6700XT out for a 1660 super seems to have fixed things but that's not a great solution because it's a much slower card, draws less power so won't tell me if it's a power issue and I need it elsewhere. I can theoretically test the 6700XT in another machine but it's PSU almost certainly isn't powerful enough and the card won't physically fit in the case.

The 6700XT is theoretically still under warranty but it was itself a warranty replacement for a Radeon VIII that failed with 2 weeks left so getting Gigabyte to accept a return might be difficult.

Recent changes: Nope, weather has got colder as well so it shouldn't be temperature related
--

Operating system: Windows 10
System specs:
Asrock B550 Phantom Gaming ITX
Ryzen 5900X
32GB (2x16) G.Skill Trident Z 3200C15
Gigabyte 6700XT Eagle
600W Silverstone Strider SST-SX600-G PSU
Dan A4 case (with PCI-E riser)

Location: UK

I have Googled and read the FAQ: Yes

Adbot
ADBOT LOVES YOU

down1nit
Jan 10, 2004

outlive your enemies
Agreed, rma while you can.

The techs doing rma are looking for checklist items. Here's my normal checklist that gets all my rmas approved:



Generally, I would start by doing a clean install on another ssd just to have a clean slate. You can skip this but they like to see it so don't, it doesn't take that long. Assuming it gets thru the install without glitching, immediately go into device manager and disable the graphics adapter so windows is using basic display drivers. This gets rid of any VGA driver potential.

Next install the amd ryzen package and run all the benchmark tools I can, including prime 95, cine bench, and hwinfo64. Hopefully it crashes and you can just list what you did. If it doesn't, enable the vga driver again and get to gaming. If/when it crashes make note and send that poo poo to amd. As long as it's not now a GPU error message...

If it doesn't crash, yay! Keep it like this for a day, for a few days or more to make sure.

Yeep
Nov 8, 2004
The other thing I guess is it could be the PSU being underpowered at 600w. Plugging everything into a couple of online calculators gets me between 500 and 650w requirement. I'm not sure why that would suddenly be an issue after 3 years though.

down1nit
Jan 10, 2004

outlive your enemies
Good detail, worth investigating too.

There are certain chemistry things in big rear end capacitors that can make power supplies less efficient over time, and that can then lead to insufficient power availability. You have at least one big rear end capacitor in there.

Yeep
Nov 8, 2004

down1nit posted:

Good detail, worth investigating too.

There are certain chemistry things in big rear end capacitors that can make power supplies less efficient over time, and that can then lead to insufficient power availability. You have at least one big rear end capacitor in there.

I re-seated the PSU end of the modular PCI-E power cable and that seems to have fixed things for now. I also remembered that I'm running a 6700XT with 1x 8pin and 1x 6 pin power off a single cable with a splitter (because there's only one on the PSU). Time to try and track down a Corsair SF750 I think.

Adbot
ADBOT LOVES YOU

down1nit
Jan 10, 2004

outlive your enemies
Noice. I agree, in general it's good to have a new PSU anyway at a certain age. Now seems right!

Good luck and bump if need

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply