Had a weird batch of behaviour with my Vega64 (reference model) the past couple of nights.
Realising there was probably some symptoms indicative of the behaviour prior but anyway, as in sometimes, the screen would flicker momentarily and id lose signal for about 2 seconds, then itd come back on, so I've been trying to figure it out but I am at a loss.
I came home from work to notice the red LED off on the GPUTach. HDMI signal was missing - e.g monitor trying to find an output. Restarted my machine, booted into windows. Loading a game, and then the graphics card Tach LED went off again and the fan stopped moving when i checked for any sort of hardware failure (radeon lights stayed on though) and the rest of my pc was still on. I rebooted my machine, it got to the windows page, then as it moved past, did the same thing again. Thinking this was a driver problem.
I rebooted into safe mode, everything worked, completely removed drivers using both the AMD driver removal tool and the Guru3D one. I tried a few games just to see if it was a specific problem e.g power consumption making my card bail and alas, no problems. This worked momentarily, then started happening again, gets past login, then it seems as soon as it boots maybe, a driver, or AMD software it does the same thing. I can't really verify that further, however, I can do some more tests.
Now, I'm writing this from Linux as I spend about 80% of my time on Linux and 20% on Windows gaming or whatever, and it works fine. I thought this was isolated to windows until i tried running Doom, which bailed on first run with the same symptoms. However, I can then reboot go into Unigine benchmarks and stress test my card, where I can visible see the power load, which has a bit baffled. Just then i was able to run Doom fine for about an hour.
I've checked my Event logs / dmesg and I can't really spot anything damning.
Id really appreciates some help into how I can diagnose this and get some consistent data.
What I've done so far:
Ive switched BIOS on my GPU
I've flashed my motherboard to latest BIOS
Gone to 2 separate PCI-e cables from the power supply rather than the 1 cable (normally had the dual connector on the 1 lead)
memtested / changed DIMMs.
Checked PSU cabling.
My Setup:
Ryzen 2700X
32GB Corsair Dominator DDR4 3200
Asus B450-F ROG Strix.
Vega 64 Reference
Seasonic Focus GX 750 750W Modular 80+ Gold PSU
I would assume if it was PSU, id know by now. This PC has been used a lot this year, and this is the first sign of it.
No overclocks. No undervolts. All Stock settings.
