Use software's BSP node traversal in OpenGL
1 unresolved thread
1 unresolved thread
I'm honestly surprised how this has gone unnoticed for this long. Turns out software has had an optimization in it's BSP tree traversal that avoids unnecessary recursion for quite a while, but it was never implemented in OpenGL, so I'm doing it now. I haven't done any performance tests, though, since I assume that whoever made the optimization in the software renderer already has.
Merge request reports
Activity
Filter activity
Nope. The original doom used what OpenGL uses now. Someone else changed the software renderer to something more efficient.
added OpenGL renderer Optimisation Rendering labels
mentioned in commit a2aaa45d
Please register or sign in to reply