Statistical model of program/verify algorithms in resistive-switching memories for in-memory neural network accelerators
- Resistive-switching random access memory (RRAM) is a promising technology for in-memory computing (IMC) to accelerate training and inference of deep neural networks (DNNs). This work presents the first physics-based statistical model describing (i) multilevel RRAM device program/verify (PV) algorithms by controlled set transition, (ii) the stochastic cycle-to-cycle (C2C) and device-to-device (D2D) variations within the array, and (iii) the impact of such imprecisions on the accuracy of DNN accelerators. The model can handle the full chain from RRAM materials/device parameters to the DNN performance, thus providing a valuable tool for device/circuit codesign of hardware DNN accelerators.
Author: | Artem Glukhov, Valerio Milo, Andrea Baroni, Nicola Lepri, Cristian Zambelli, Piero Olivo, Eduardo Perez, Christian WengerORCiD, Daniele Ielmini |
---|---|
DOI: | https://doi.org/10.1109/IRPS48227.2022.9764497 |
ISBN: | 978-1-6654-7950-9 |
ISBN: | 978-1-6654-7951-6 |
ISSN: | 2473-2001 |
Title of the source (English): | 2022 IEEE International Reliability Physics Symposium (IRPS) |
Publisher: | Institute of Electrical and Electronics Engineers (IEEE) |
Document Type: | Conference publication peer-reviewed |
Language: | English |
Year of publication: | 2022 |
Tag: | Multilevel switching; RRAM; neural network |
First Page: | 3C.3-1 |
Last Page: | 3C.3-7 |
Faculty/Chair: | Fakultät 1 MINT - Mathematik, Informatik, Physik, Elektro- und Informationstechnik / FG Halbleitermaterialien |