A 22-nm 1,287-MOPS/W Structured Data-Path Array for Binary Ring-LWE PQC

Proceedings of ESSCIRC, pp. 189–192

Abstract

A $0.0131-\mathrm{mm}^{2}$ structured data-path array for binary Ring-LWE (BRLWE) post-quantum cryptography (PQC) is fabricated in 22-nm CMOS with an energy efficiency of 1,287 MOPS/W at $0.45 \mathrm{~V}$ and throughput of 6.92 MOPS at $0.8 \mathrm{~V}$. The proposed digital compute-in-memory (DCIM) architecture maximizes the data locality, introduces a novel ternary multiplication method that halves the arithmetic complexity, and optimizes the negative wrapped convolution scheme resulting in a structured dataflow of $23 \%$ smaller area, 2.6X higher throughput and 2X better energy efficiency than SotA. Furthermore, the proposed DCIM architecture significantly reduces the information leakage of cache and timing channels by statically storing the key words in memory.

Authors 2

1. Shutao Zhang Aachen Chair of Integrated Digital Systems and Circuit Design (IDS)

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Chair of Integrated Digital Systems and Circuit Design,Germany

   Chair of Integrated Digital Systems and Circuit Design, RWTH Aachen University, Germany

2. Tobias Gemmeke Aachen Chair of Integrated Digital Systems and Circuit Design (IDS)

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Chair of Integrated Digital Systems and Circuit Design,Germany

   Chair of Integrated Digital Systems and Circuit Design, RWTH Aachen University, Germany