Optimizing BIKE for the Intel Haswell and ARM Cortex-M4

BIKE is a key encapsulation mechanism that entered the third round of NIST post-quantum cryptography standardization process. This paper presents two constant-time implementations for BIKE, one tailored Intel Haswell and ARM Cortex-M4. Our implementation much faster than avx2 written by team: bikel1, level-1 parameter set, we achieve 1.39x speedup decapsulation (which slowest operation) 1.33x sum all operations. For bikel3, level-3 1.5x 1.46x M4 more times non-constant-time portable team. The speedups are achieved both algorithm-level instruction-level optimizations.