PIMA-LPN: Processing-in-memory Acceleration for Efficient LPN-based Post-Quantum Cryptography
DescriptionLearning parity with noise (LPN) has recently attracted intensive research for building efficient quantum-secure multiparty computation protocols. In this paper, we design PIMA-LPN, a processing-in-memory (PIM) accelerator for LPN cryptography. Specifically, since PIM can handle NAND much better than AND, we propose a NAND-substitute procedure for computing LPN. Rigorous experiments demonstrate that PIMA-LPN is 20.86x (resp. 25.5x) and 216.8x (resp. 143.3x) faster than the FPGA and CPU implementations of LPN encryption (resp. decryption) functions guaranteeing 256-bit of security, respectively.
Event Type
Research Manuscript
TimeTuesday, July 11th4:10pm - 4:25pm PDT
Location3012, 3rd Floor
Hardware Security: Primitives, Architecture, Design & Test