Answers are grounded in this paper and cite their sources.
Modular division is an essential but time-consuming computation in elliptic curve cryptography (ECC). This paper presents evaluation results of polynomial modular division algorithms over Galois binary finite fields using open-source and reproducible Verilog HDL code, with a particular focus on both Fermat’s Little Theorem (FLT) using the Itoh-Tsujii algorithm (ITA) and the extended Euclidean algorithm (EEA) in the same environment. To reduce the execution time of polynomial modular division, we investigate several architectures of ITA that use parallel squarers (PSQs). To reduce the kilo-area-time (KAT) product, we propose a polynomial Brunner extended Euclidean division algorithm (BEEDA) and a radix-4 BEEDA that achieve a very high frequency in hardware implementations. The experimental results on an Altera Cyclone 10 FPGA chip show that PSQs with 9, 29, and 3 squarers achieve shorter division execution time of 0.65 μs, 0.73 μs, and 1.97 μs; the BEEDA achieves lower KATs of 2.75, 5.72, and 9.11; and the radix-4 BEEDA achieves lower KATs of 2.51, 5.20, and 7.82, for Koblitz curves K-163, K-233, and K-283, respectively. Based on this evaluation, we conclude that PSQs are suitable for platforms with high-speed requirements and abundant resources, while our proposed BEEDA or radix-4 BEEDA is the optimal choice for resource-constrained FPGAs.
No takes yet. Share an insight, caveat, or question.
Yamin Li (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: