The Impact of Allocating Weights and Biases in SRAM Parity Checking Areas on the Soft Error Reliability of CNN Inference Models
Year: 2024
Type: conference-paper
Abstract: This paper promotes a novel semi-automated compiler-based technique that allocates CNN weights and bias to SRAM parity checking areas. Radiation results suggest that it is possible to reduce critical SDC events by up to 90%.
Authors: J. Gava, L. Laurini, R. Possamai Bastos, F. Moraes, R. Reis, L. Ost
Institutions: Universidade Federal do Rio Grande do Sul, Institut polytechnique de Grenoble, Centre National de la Recherche Scientifique, Techniques of Informatics and Microelectronics for Integrated Systems Architecture, Pontifícia Universidade Católica do Rio Grande do Sul, Loughborough University
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4399793726