Experimental Release 2025.11.12
https://downloads.luoshuangfw.io/gpulm-releases/GPULM-5.6.1.7z
Back to release list
With the GPU denoiser being 10-15x faster, now it’s time to optimize the lightmap transcoding and encoding part. It was not worth doing as the CPU denoiser was taking the majority of time.
This release optimizes by parallelization. Though, processing many meshes simultaneously can take lots of memory. Let me know if you run into memory issues.