Year: 2022
Type: conference-paper
Abstract: The precise programming of crossbar arrays of unit-cells is crucial for obtaining high matrix-vector-multiplication (MVM) accuracy in analog in-memory computing (AIMC) cores. We propose a radically different approach based on directly minimizing the MVM error using gradient descent with synthetic random input data. Our method significantly reduces the MVM error compared with conventional unit-cell by unit-cell iterative programming. It also eliminates the need for high-resolution analog-to-digital converters (ADCs) to read the small unit-cell conductance during programming. Our method improves the experimental inference accuracy of ResNet-9 implemented on two phase-change memory (PCM)-based AIMC cores by 1.26%.
Authors: J. Büchel, A. Vasilopoulos, B. Kersting, F. Odermatt, K. Brew, I. Ok, S. Choi, I. Saraf, V. Chan, T. Philip, N. Saulnier, V. Narayanan, M. Le Gallo, A. Sebastian
Institutions: IBM Research - Zurich, IBM Research - Albany, IBM (United States)
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4317793291
