Return to Answer

This one was harder for me to wrap my head around, but my ascii art of the masks seem to make sense.

This one was harder for me to wrap my head around, but my ascii art of the masks seems to make sense.

There is much literature on this subject but it may not be immediately obvious that the results are applicable here. A bit-cube can be considered as a stack of bit-planes. And rotation of a bitmap is a topic that underwent much study at Xerox PARC when they invented WIMP and GUI. An algorithm is presented in the [Byte Magazine 1981 SmallTalk Issue] (https://archive.org/details/byte-magazine-1981-08) on page 188.

There is literature on this subject but it may not be immediately obvious that the results are applicable here. A bit-cube can be considered as a stack of bit-planes. And rotation of a bitmap is a topic that underwent much study at Xerox PARC when they invented WIMP and GUI. An algorithm is presented in the [Byte Magazine 1981 SmallTalk Issue] (https://archive.org/details/byte-magazine-1981-08) on page 188.

A naive counting of the delta-swap's operations is 1 AND, 2 SHIFTs and 3 XORs which if called 4 times yields 4 ANDs, 8 SHIFTs, and 12 XORs for the whole rotation. So the optimum may well depend upon the particular instructions' speeds on a particular hardware implementation. More ANDs and less (X)ORs or vice versa?