Triple monitor setup with 2x Acer K242HL (1080, 60hz) and an AOC G2460PF (1080, 144hz) 5 fan setup
I read beforehand that the R9 390 was quite power-consuming, but I have never had issues with it so it ended up not mattering.
With the relase of AMD's Ryzen series, I switched my mobo to ASUS Prime B350-PLUS, got a Ryzen 5 1600x and upgraded to Windows 10 near the end of 2017.
Both setups worked perfectly and I never had any problems until now.
Yesterday something weird happened while playing World of Warcraft, where the game crashed and all 3 monitors started flickering. I tried rebooting, but it kept happening and starting a game (Apex Legends, WoW, BF1, DOOM) increased the amount and frequency of the flickering, as well as causing all monitors to turn off when the GPU was under heavy load. When it first occurred, I was using Radeon Adrenalin 19.1.1, and updating to the optional 19.2.3 did nothing. All temps are fine, tested with AMD Wattman, MSI Afterburner, Speccy, BIOS and HWMonitor.
I've been trying to find information and fixes, and so far the things I've tried are:
- AMD's DDU or Guru3D's DDU (in safe mode) to remove all GPU software and trying versions 19.1.1, 18.12.3, 18.12.2, 18.10.2, 18.2.3, 17.1.1, 16.2.1, 15.7.1, 15.6, 15.5
15.5, 15.6 and 15.7.1 were not installed properly and caused my GPU to be detected as R9 200, and I read in an AMD thread that it's due to incompatibility with older drivers. None of the software versions I tried made any difference, although I isolated the problem specifically to the AMD display adaptor and not the Radeon Settings install. Disabling the AMD display adaptor in device manager stops the corruption from occurring. The corruption follows my GPU if i connect it to another system with the AMD display adaptor installed (with different monitor), and ceases if the AMD display adaptor is disabled.
My monitors use different cables, one DVI-D, one HDMI-DVI-D adapter, and one DP. All 3 cables work without issues on another system. On my current system, the display issue occurs regardless of which monitor(s) is connected with which cable. I have no issues with black screens aside from when the GPU is under heavy load.
- Clean windows install with the latest AMD drivers
- Uninstalling the display adaptor incl. software through device manager, rebooting and letting windows download the AMD display driver
- AMD's DDU (in safe mode) and using windows update to find AMD drivers
- Cleaning the PC and reseating the GPU
- Unplugging the power cord for 10 min
- Holding the power button for 40s
- Booting with 1, 2 and 3 monitors connected at a time
- Booting with no monitor connected and only connecting one with all cable combinations
- Setting the 2D memory clock to match 3D
- Updating BIOS from 3203 to 4023
- Disabling Radeon Wattman controls
- Disabling MSI Afterburner controls
- Disabling ULPS in MSI Afterburner
- Setting an aggressive GPU fan curve in MSI Afterburner/Wattman
- Increasing GPU power limit in MSI Afterburner/Wattman
- Lowering GPU clock
- Lowering memory clock
- Forcing higher clock rates
- Turning VSR on/off
- Changing monitor refresh rate and/or resolution
- Enable/disable eyefinity
- Monitoring memory clock when the corruption occurs (no weird spikes or forced low clocks)
- Running one monitor in higher resolution and 1 and/or 2 in lower resolution
- Running DISM
- Running SFC
- Uninstalling/reinstalling Microsoft Visual C++ Redistributable
To sum things up:
The AMD display adaptor is causing display corruption with my R9 390, and disabling it removes the corruption but is not a solution. Tried multiple monitors/system with the same result. No temperature issues.