SAMO: Optimised Mapping of Convolutional Neural Networks to Streaming Architectures

TL;DR

SAMO is a framework that optimizes the mapping of CNNs onto Streaming Architectures in FPGAs, significantly improving performance by exploring the design space and FPGA reconfigurability.

Contribution

This work introduces SAMO, a unified optimization framework that enhances CNN-to-FPGA streaming architecture mappings by exploiting model structure and FPGA reconfigurability.

Findings

SAMO achieves 4-20x better performance than existing designs.

Three optimization methods enable flexible, high-performance mappings.

The framework is open-source for community use and development.

Abstract

Significant effort has been placed on the development of toolflows that map Convolutional Neural Network (CNN) models to Field Programmable Gate Arrays (FPGAs) with the aim of automating the production of high performing designs for a diverse set of applications. However, within these toolflows, the problem of finding an optimal mapping is often overlooked, with the expectation that the end user will tune their generated hardware for their desired platform. This is particularly prominent within Streaming Architecture toolflows, where there is a large design space to explore . In this work, we establish the framework SAMO: a Streaming Architecture Mapping Optimiser. SAMO exploits the structure of CNN models and the common features that exist in Streaming Architectures, and casts the mapping optimisation problem under a unified methodology. Furthermore, SAMO explicitly explores the reconfigurability property of FPGAs, allowing the methodology to overcome mapping limitations imposed by certain toolflows under resource-constrained scenarios, as well as improve on the achievable throughput. Three optimisation methods - Brute-Force, Simulated Annealing and Rule-Based - have been developed in order to generate valid, high performance designs for a range of target platforms and CNN models. Results show that SAMO-optimised designs can achieve 4-20x better performance compared to existing hand-tuned designs. The SAMO framework is open-source: https://github.com/AlexMontgomerie/samo.