DeepAI AI Chat
Log In Sign Up

Unified Characterization Platform for Emerging NVM Technology: Neural Network Application Benchmarking Using off-the-shelf NVM Chips

by   Supriya Chakraborty, et al.
Indian Institute of Technology Delhi

In this paper, we present a unified FPGA based electrical test-bench for characterizing different emerging NonVolatile Memory (NVM) chips. In particular, we present detailed electrical characterization and benchmarking of multiple commercially available, off-the-shelf, NVM chips viz.: MRAM, FeRAM, CBRAM, and ReRAM. We investigate important NVM parameters such as: (i) current consumption patterns, (ii) endurance, and (iii) error characterization. The proposed FPGA based testbench is then utilized for a Proof-of-Concept (PoC) Neural Network (NN) image classification application. Four emerging NVM chips are benchmarked against standard SRAM and Flash technology for the AI application as active weight memory during inference mode.


Benchmarking High Bandwidth Memory on FPGAs

FPGAs are starting to be enhanced with High Bandwidth Memory (HBM) as a ...

SoftMC: Practical DRAM Characterization Using an FPGA-Based Infrastructure

This paper summarizes the SoftMC DRAM characterization infrastructure, w...

FPGA-based Hyrbid Memory Emulation System

Hybrid memory systems, comprised of emerging non-volatile memory (NVM) a...

A Highly Parallel FPGA Implementation of Sparse Neural Network Training

We demonstrate an FPGA implementation of a parallel and reconfigurable a...

Low-Rank Training of Deep Neural Networks for Emerging Memory Technology

The recent success of neural networks for solving difficult decision tal...

WURBench: Toward Benchmarking Wake-up Radio-based Systems

The performance of wake-up radios must be clearly measured and understoo...