Binary Neural Networks in Spintronic Memory

Abstract

Neural networks span a wide range of applications of industrial and commercial significance. Binary neural networks (BNN) are particularly effective in trading accuracy for performance, energy efficiency or hardware/software complexity. In this thesis, I demonstrate a spintronic, re-configurable in-memory BNN accelerator, affectionately named PIMBALL: Processing In Memory BNN Accelerator, which allows for massively parallel and energy efficient computation. PIMBALL is capable of being used as a standard spintronic memory (STT-MRAM) array and a computational substrate simultaneously. For evaluation, I test PIMBALL using multiple image classifiers and a genomics kernel. Thesimulation results show that PIMBALL is more energy efficient than alternative CPU, GPU, and FPGA based implementations while delivering higher throughput.