Key generation creates the top Merkle tree; if we cache parts of it, it speeds up signature generation.
This is actually a fairly minor speed up for the standard SLH-DSA parameter sets; it is a significant improvement to the _24 parameter sets.
Should we implement it?
Key generation creates the top Merkle tree; if we cache parts of it, it speeds up signature generation.
This is actually a fairly minor speed up for the standard SLH-DSA parameter sets; it is a significant improvement to the _24 parameter sets.
Should we implement it?