|
|
|
|
|
|
by trishume
1798 days ago
|
|
|
This is awesome. In theory you could absolutely minimize the latency penalty to just the overhead of the gpu1->memory->gpu2 copy, if the display sync signals from the display the passthrough window was on were passed through to the GPU driver on Windows, and that was combined with fullscreen compositor bypass (available on many Linux WMs) or low-latency compositing (available on sway and now mutter https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1762 on Wayland). |
|
|