The number of threads that are executed simultaneously by the GPU.
- iOS 8.0+
- macOS 10.11+
- tvOS 9.0+
For better performance, when dispatching a compute command using the
MTLCompute, make the number of threads in the threadgroup a multiple of
thread. If you can guarantee that you will always dispatch compute commands in this way, use a
MTLCompute object to create the pipeline state object, and set the descriptor's
thread property to true.