This makes the n * m pattern more efficient (w-NAF tables / forms are only computed once for each base / scalar) and has more predictable memory usage (one w-NAF table of the specified window size per base, one w-NAF form of the specified window size per scalar).
This makes the
n * mpattern more efficient (w-NAF tables / forms are only computed once for each base / scalar) and has more predictable memory usage (one w-NAF table of the specified window size per base, one w-NAF form of the specified window size per scalar).