Hello,
I'm reporting an issue related to GPU process management when launching Vortex on newer AMD hardware (tested on a RX 7900 XT, drivers fully updated, Windows 11).
Without any special flags, Vortex fails to start: the Electron GPU process repeatedly crashes with exit code -1073741515, resulting in a fatal error:
[ERROR:gpu_process_host.cc(993)] GPU process exited unexpectedly: exit_code=-1073741515 [FATAL:gpu_data_manager_impl_private.cc(448)] GPU process isn't usable. Goodbye.
The workaround is to launch Vortex with the --in-process-gpu flag:
Vortex.exe --in-process-gpu
With this flag, the GPU code runs in the main process instead of a separate GPU process, and Vortex starts and works correctly.
Some observations:
The crash happens extremely early, before UI rendering begins.
Electron logs indicate a failure to create a GLES3 context, followed by a fallback attempt to GLES2, and ultimately a fatal failure to initialize any GPU context.
This behavior is typical when an Electron version is not fully compatible with modern GPU driver models (e.g., WDDM 3.x, Shader Model 6.x).
I also noticed deprecation warnings about certain Electron APIs (e.g., crashed event vs render-process-gone), suggesting the Electron version in use may be outdated relative to current hardware and driver expectations.
Do i have to change something ? And will it cause problem using the --in-process-gpu flag ?
Ty !