Open
Description
Describe the feature you'd like to request
Recognize to launch preview generation for next batch of GPU based processing at the time of launching GPU processing.
Describe the solution you'd like
CPU used to generate previews whilst GPU processes previews from previous CPU batch.
It would double the speed of processing vs current linear approach of: generate previews (CPU used), once finished, pushed to GPU processing. At the same time, background process could be launched to generate previews required for next run.
Describe alternatives you've considered
Current implementation spends half if not more of the time generating previews using CPU whilst GPU is left unused and vice versa when previews are generated and GPU kicks in, CPU is relatively unused.