This paper presents a comprehensive study on the implementation of a RISC-V-based memory-mapped accelerator designed for Number Theoretic Transform (NTT) and Inverse Number Theoretic Transform (INTT) operations within the context of the post-quantum cryptographic algorithm CRYSTALS-Kyber. The primary focus lies in the performance evaluation of the algorithm, with a particular emphasis on minimizing the overhead associated with transferring data between the core and the implemented IP. The analysis includes a deep dive into the intricacies of data transfer, leveraging Direct Memory Access (DMA) to efficiently reduce overhead. The evaluation results show that our approach, when applied to a X-HEEP core, achieves up to 15.7x and 19.6x improvement in cycle count for NTT and INTT respectively, compared to the base software implementation. To this end, we also demonstrate the efficacy of the proposed memory-mapped accelerator in enhancing the overall performance of CRYSTALS-Kyber, thereby contributing to the advancement of secure cryptographic systems in the post-quantum era.
Implementation and integration of NTT/INTT accelerator on RISC-V for CRYSTALS-Kyber / Dolmeta, Alessandra; Valpreda, Emanuele; Martina, Maurizio; Masera, Guido. - ELETTRONICO. - 1:(2024), pp. 59-62. (Intervento presentato al convegno Proceedings of the 21st ACM International Conference on Computing Frontiers Workshops and Special Sessions tenutosi a Ischia (Italy) nel May 7-9, 2024) [10.1145/3637543.3652872].
Implementation and integration of NTT/INTT accelerator on RISC-V for CRYSTALS-Kyber
Dolmeta, Alessandra;Valpreda, Emanuele;Martina, Maurizio;Masera, Guido
2024
Abstract
This paper presents a comprehensive study on the implementation of a RISC-V-based memory-mapped accelerator designed for Number Theoretic Transform (NTT) and Inverse Number Theoretic Transform (INTT) operations within the context of the post-quantum cryptographic algorithm CRYSTALS-Kyber. The primary focus lies in the performance evaluation of the algorithm, with a particular emphasis on minimizing the overhead associated with transferring data between the core and the implemented IP. The analysis includes a deep dive into the intricacies of data transfer, leveraging Direct Memory Access (DMA) to efficiently reduce overhead. The evaluation results show that our approach, when applied to a X-HEEP core, achieves up to 15.7x and 19.6x improvement in cycle count for NTT and INTT respectively, compared to the base software implementation. To this end, we also demonstrate the efficacy of the proposed memory-mapped accelerator in enhancing the overall performance of CRYSTALS-Kyber, thereby contributing to the advancement of secure cryptographic systems in the post-quantum era.File | Dimensione | Formato | |
---|---|---|---|
3637543.3652872.pdf
non disponibili
Tipologia:
2a Post-print versione editoriale / Version of Record
Licenza:
Non Pubblico - Accesso privato/ristretto
Dimensione
490.88 kB
Formato
Adobe PDF
|
490.88 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
ACM_OSHW2.pdf
accesso aperto
Tipologia:
2. Post-print / Author's Accepted Manuscript
Licenza:
PUBBLICO - Tutti i diritti riservati
Dimensione
448.52 kB
Formato
Adobe PDF
|
448.52 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/11583/2990228