It should be possible to split big volume into smaller chunks, do transformation on GPU and then combine the chunks on CPU
It should be possible to split big volume into smaller chunks, do transformation on GPU and then combine the chunks on CPU