Improved the Automated Evaluation Algorithm Against Differential Attacks and Its Application to WARP

This paper presents a heuristic approach to searching the key recovery-friendly distinguishers for block ciphers, which aims to attack more rounds with lower complexities. Firstly, we construct an SAT model to search for a set of distinguishers with the minimum number of active input-output words (and optimal probability). Subsequently, based on the discovered distinguishers, we select the advantageous distinguisher with fewer key bits involved in the key recovery phase. Finally, the guess-and-check for the key recovery attack is performed using the manual approach to compute the attack parameters accurately. By applying our new technique to \(\texttt{WARP}\) proposed in SAC 2020, we identify some 19-round and 20-round advantageous differentials. Simultaneously, the high-probability chain of Sbox leads to a stronger clustering effect of the differential trails for \(\texttt{WARP}\) , so we effectively improve the probability of the advantageous distinguisher. Also, the first 25-round differential attacks are performed by extending a 19-round distinguisher and a 20-round distinguisher, respectively. The results cover 2 more rounds than the previous known differential attacks.