Yeah, that's a classic "your GPU is dying" scenario. DirectX errors + crashes specifically when overclocked = the silicon can't handle the workload anymore.
The OC is just the stress test that exposes the problem - it might run fine at stock clocks for a bit longer, but it's on borrowed time. Could be:
- VRAM degradation (most common with heavily used cards)
- Core instability from years of thermal cycling
- Power delivery issues on the card itself
My advice:
- Immediate: Drop any OC, run at stock or even underclock slightly - might buy you a few weeks/months
- Test: Run stock clocks and see if crashes stop - if they do, confirms it's the card struggling
- Monitor temps: Could be thermal paste dried out, but if temps are fine and it still crashes, it's component failure
If it crashes even at stock clocks on BF6 (which is demanding but not unreasonable), the card's cooked. Time to start shopping.
Here are some stock solutions that might work:
Drivers:
- DDU clean install in safe mode
- Update to latest WHQL drivers
- Try previous stable driver version
- AMD: avoid beta drivers, use 24.x.x series
- Nvidia: try 566.03, 561.09, or 560.94
GPU Settings:
- Remove ALL overclocks (core + memory)
- Reset to stock clocks
- Underclock by 50-100 MHz if still crashing
- Undervolt instead of overclock
- Increase power limit to max available
- Lower power limit if overheating
- Disable auto-tuning/auto-OC features
In-Game Settings:
- Lower texture quality
- Reduce mesh quality
- Disable ray tracing
- Disable DLSS/FSR/frame generation
- Turn off dynamic resolution
- Lower overall graphics preset
- Cap framerate (60fps or monitor refresh rate)
- Reduce resolution
Windows Settings:
- Run game as administrator
- Disable fullscreen optimizations (game .exe properties)
- Set game to high performance mode in Graphics Settings
- Verify using discrete GPU, not integrated graphics
- Disable hardware-accelerated GPU scheduling
- Disable Windows Game Mode
- Reinstall DirectX runtime
- Run: sfc /scannow
- Run: DISM /Online /Cleanup-Image /RestoreHealth
Software Conflicts:
- Close ALL overlays (Discord, Steam, GeForce Experience, AMD overlay)
- Disable MSI Afterburner, RTSS, HWInfo
- Exit RGB control software
- Disable antivirus during gaming
- Close browser/streaming apps
GPU-Specific:
- Clear shader cache (AMD Adrenalin or Nvidia Control Panel)
- Nvidia: disable G-Sync
- AMD: disable anti-lag/anti-lag+
- Set prefer maximum performance in control panel
- Disable power saving features
Game-Specific:
- Verify/repair game files
- Switch DX12/DX11 if available
- Delete config files and reset settings
- Reinstall game
Hardware Checks:
- Monitor GPU temps (shouldn't exceed 83°C)
- Check PSU quality (650W+ recommended, quality brand)
- Reseat GPU in PCIe slot
- Check power cables fully seated (especially 12VHPWR/8-pin)
- Update motherboard BIOS
- Check RAM with MemTest86
- Monitor VRAM usage (shouldn't max out)
Diagnostics:
- Run 3DMark Time Spy stress test
- Run FurMark for 10 minutes
- Test different games
- Check Windows Event Viewer for details
- Monitor with GPU-Z during crash
If Still Crashing:
- At stock in all games = RMA/replace GPU
- Only in one game = game bug/incompatibility
- Only when OCed = silicon limit, run stock
- Only under heavy load = thermal/power issue
