This is still on-going for thousands of players EA, can we get an update?
Happened twice during the first 35 hours of play, that I can live with.
Then 5 times over 4 hours.
Now it's every 2-20 minutes. I can't play the game anymore. I can't finish a match.
Tried every solution Ive come across so far, reinstalled the game, direct X, swapped drivers,
installed/uninstalled windows updates, resizable bar on/off, DCOP/XMP on/off, PBO fiddling and on/off, page file sizes, shader caches both for the game and dx, all in-game settings wiped, from off/low to overkill etc. etc.
Im at my wits end here and I put in a refund request to Steam and I suggest others do as well until this is fixed since the game is just broken on a technical level at the moment.
Τhe game consistently crashes after 5–15 minutes of play, no matter what settings I use.
The issue started on the second day after launch — on day one I could play for ~10 hours straight without a single crash. Since then, the game crashes every session with either:
"DirectX function GetDeviceRemovedReason failed with DXGI_ERROR_DEVICE_HUNG", or
Access Violation (0xC0000005) pointing to RenderCore2.PlatformPcDx12.dll or RenderCore2.PlatformPcDx11.dll.
I have multiple .mdmp crash dumps confirming that the crash happens inside Frostbite RenderCore2 — specifically during GPU command submission, not in the driver layer.
System Specs
CPU: AMD Ryzen 9 9800X3D
GPU: NVIDIA GeForce RTX 3080 (EVGA XC3)
Motherboard: MSI B650 Tomahawk MAX
RAM: 32 GB DDR5 6000 MHz CL30
PSU: Lian Li 1000 W Gold
Storage: NVMe Gen4 SSD
OS: Windows 11 Pro 64-bit, fully updated
GPU Driver: NVIDIA 552.44 WHQL (tried 581.57 and 537.58 too)
Monitor: 1440p / 240 Hz G-Sync
Error Details from Crash Dumps
DX12 mode: DXGI_ERROR_DEVICE_HUNG inside
RenderCore2.PlatformPcDx12.dll → D3D12Core.dll → nvldumdx.dll.
DX11 mode: Access Violation (0xC0000005) inside
RenderCore2.PlatformPcDx11.dll → D3D11.dll.
DRED data shows:
AllocationName: Frostbite::Render::DX12CommandQueue
→ Indicates the engine lost its GPU device context or invalidated its command buffer.
No overlays, overclocks, or third-party DLL hooks active.
Crash address is always the same (0x...66CA), confirming deterministic bug inside RenderCore2.
Troubleshooting Already Tried
I’ve already tested every known fix and still get the same result:
Drivers & Windows
- Clean driver install with DDU in Safe Mode.
- Tried 581.08, 552.44 WHQL, 581.57 and 537.58 WHQL (same issue).
- Windows fully updated, GPU firmware up to date.
System stability
- No overclock on CPU/GPU/RAM (tested also with –50 MHz core / –200 MHz memory).
- GPU temps under 60 °C, VRAM < 60 °C.
- MemTest64 and OCCT show 0 errors.
- PSU 1000 W Gold, stable voltages.
Settings & Engine
- Disabled HAGS, Game Mode, ReBAR, iGPU.
- Tried both DX12 and DX11 (same crash).
- Reset and deleted all game config and shader cache folders.
- Disabled Ray Tracing, DLSS, Reflex Boost, On-Demand Texture Streaming.
- Capped FPS to 165 Hz, V-Sync Off.
- Disabled all overlays (EA, Steam, Discord, Afterburner, RTSS).
- Attempted -RenderDevice.RenderAheadLimit 1 and -RenderDevice.Dx12Enable 0.
- Disabled Windows MPO via registry.
Hardware isolation
- Tested with undervolt (0.900 V @ 1850 MHz) and stock — identical results.
- No other games crash (COD, Cyberpunk, Valorant, etc. all stable).
All crash dumps show the same pattern:
Frostbite RenderCore2 triggers a GPU device hang or memory access violation inside the DX12/DX11 command queue layer.
This strongly suggests a regression introduced with patch v2.2.0 affecting RTX 30-series GPUs and X3D CPUs (high frame-time variance).
It’s not caused by drivers or hardware — it’s an internal synchronization or memory-corruption bug inside the RenderCore2 module.
I attach a Google Drive folder with some of my crashes:
https://drive.google.com/drive/folders/1bUSGO48XkK-4WSGIsDYzd87tw1p-nAoF?usp=sharing
