See about Getting OCIO Transforms on GPU

[AltaArts]

Right now the pipeline is:

• Caching into RAM on CPU (has to be for formats such as ProRes)
• RAM - > CPU for OCIO calcs -> RAM
• RAM -> GPU for Display

this is working pretty well and fast, but will explore creating the OCIO transform in a GPU shader so it will just go:

RAM -> GPU (OCIO) -> GPU (Display)

[AltaArts]

Yeah, so I have not been able to get this to work for tranforms that need LUTs. Right now the CPU based OCIO transforms seem to be plenty fast, so maybe just leave it.