We wan to be able to select from three different probability models:
Cell-wise Probability propagation
Column-wise Probability propagation
Fullstate Probability propagation
It should also be possible to mix-and-match the different models throughout the cipher, since e.g. Fullstate is the most accurate, but unfeasible to calculate for states with many possible state differences.
We wan to be able to select from three different probability models:
It should also be possible to mix-and-match the different models throughout the cipher, since e.g. Fullstate is the most accurate, but unfeasible to calculate for states with many possible state differences.