A preview is not available for this record, please engage by choosing from the available options ‘download’ or ‘view’ to engage with the material
Description
The most commonly utilized public key cryptosystems (RSA, ECDSA, etc.) in the pre-quantum era are constructed over finite fields. For these finite fields, the most important and time-consuming operation is the modular multiplication of integers under large moduli. Efficient implementations of multi-precision arithmetic are very critical for the performance of these cryptosystems. Intel® Advanced Vector Extensions 512 (Intel® AVX-512) based Integer Fused Multiply Add Instructions (IFMA) are utilized for multi-buffer high-throughput software implementations of RSA. This paper presents a novel Modular Multiplication algorithm that increases the throughput of multi-buffer IFMA implementations of RSA operations in the range of 10%.