636 B
636 B
Old way (400 rounds, no matrix mult): 38.01s 400 was the minimum for good diffusion
New way (10 rounds, with matrix mult): 1.16s 10 rounds now give sufficient diffusion still using bitsets and strings
With new block class (instead of bitsets) (10 rounds): 0.35s still partially using bitsets and strings
With new block class (6 rounds, still good diffusion): 0.21s
With new templated block class (full block and half block): 0.14s finally no more bitsets
With new sbox, reduction, and expansion function: 0.03s
With additional jumbling in feistel rounds: 0.03s
With direct file reading/writing from data blocks: 0.02s