Applied Sciences (Sep 2024)

Optimizing HAWK Signature Scheme Performance on ARMv8

  • Siwoo Eum,
  • Minwoo Lee,
  • Hwajeong Seo

DOI
https://doi.org/10.3390/app14198647
Journal volume & issue
Vol. 14, no. 19
p. 8647

Abstract

Read online

This study proposes an optimized implementation of the HAWK Signature algorithm, one of the candidates in the first evaluation round for additional digital signature schemes in the NIST Post-Quantum Cryptography competition. The core motivation of this research is to improve the performance of HAWK algorithm. By conducting profiling analysis to identify, we identified the most resource-intensive functions. And then we optimized the functions. The optimization techniques through profiling analysis are not limited to HAWK but can be applied to other algorithms as well. Additionally, the study demonstrates how efficient optimization can be achieved using fewer instructions by leveraging lesser-known ARMv8 instructions. By targeting the functions with the highest overhead and utilizing fewer instructions, a performance improvement of approximately 2.5% for Hawk512 and 4% for Hawk1024 was achieved, respectively. These results confirm that combining profiling analysis with efficient instruction usage can lead to significant performance improvements.

Keywords