The MRSimulation file in the MolecularRenderer module uses Metal for GPU-accelerated encoding/decoding and Apple's Compression library for multicore LZBITMAP compression. The latter is discussed in #2, but can be solved with an (unoptimized) solution right now, just copying the code in the linked repo. For GPU, we would need to use some #if os gating around any usage of Metal APIs (MTLBuffer, MTLCommandBuffer, etc.) and add another compile path for using OpenCL. Then gate out all of the rendering code in the MolecularRenderer module.
The MRSimulation file in the MolecularRenderer module uses
Metal
for GPU-accelerated encoding/decoding and Apple'sCompression
library for multicore LZBITMAP compression. The latter is discussed in #2, but can be solved with an (unoptimized) solution right now, just copying the code in the linked repo. For GPU, we would need to use some#if os
gating around any usage ofMetal
APIs (MTLBuffer
,MTLCommandBuffer
, etc.) and add another compile path for using OpenCL. Then gate out all of the rendering code in the MolecularRenderer module.