DiGamma: Domain-aware Genetic Algorithm for HW-Mapping Co-optimization for DNN Accelerators

TL;DR

DiGamma is a domain-aware genetic algorithm designed for co-optimizing hardware resource configuration and mapping strategies in DNN accelerators, significantly improving search efficiency and performance over existing methods.

Contribution

It introduces a novel co-optimization framework with an efficient encoding and specialized genetic operators tailored for DNN accelerator design.

Findings

Achieves 3.0x speedup in edge settings

Achieves 10.0x speedup in cloud settings

Outperforms baseline optimization algorithms in DNN accelerator design

Abstract

The design of DNN accelerators includes two key parts: HW resource configuration and mapping strategy. Intensive research has been conducted to optimize each of them independently. Unfortunately, optimizing for both together is extremely challenging due to the extremely large cross-coupled search space. To address this, in this paper, we propose a HW-Mapping co-optimization framework, an efficient encoding of the immense design space constructed by HW and Mapping, and a domain-aware genetic algorithm, named DiGamma, with specialized operators for improving search efficiency. We evaluate DiGamma with seven popular DNNs models with different properties. Our evaluations show DiGamma can achieve (geomean) 3.0x and 10.0x speedup, comparing to the best-performing baseline optimization algorithms, in edge and cloud settings.