Answers are grounded in this paper and cite their sources.
Quantum computing poses a major threat to traditional cryptographic systems, necessitating quantum-resistant alternatives. Lattice-based cryptography, particularly NTRU lattices, offers strong security against quantum threats through the NTRU trapdoor basis. Schemes such as FALCON, BAT, and HAWK rely on efficient polynomial arithmetic, often using 64-bit fixed-point FFT to solve the NTRU equation during key generation. This work presents the first implementation of a 64-bit fixed-point FFT/IFFT architecture optimized for hardware acceleration in post-quantum cryptography. Tailored specifically for NTRU lattices, the design enhances computational efficiency through two memory-level optimizations: a conflict-free memory scheduling strategy to eliminate access contention across processing elements, and a twiddle factor compression technique that leverages symmetry and periodicity. In addition, a Karatsuba-like method is adopted for complex multiplication, reducing the number of real multiplications from four to three, significantly lowering arithmetic overhead. The proposed processor targets ASIC implementation using a 22 nm GlobalFoundries process and achieves a 400 MHz operating frequency. It achieves a 33% reduction in area and 32.5% lower power consumption than the best reported design, demonstrating its practicality for energy-efficient, high-performance post-quantum cryptographic applications.
No takes yet. Share an insight, caveat, or question.
Alshehhi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: