Bridging the Gap Between Domain-specific Frameworks and Multiple Hardware Devices

TL;DR

This paper introduces a systematic methodology that bridges domain-specific frameworks and multiple hardware devices, significantly reducing porting complexity and improving performance across diverse platforms.

Contribution

It presents a multi-layer abstraction approach that transforms domain-specific applications into hardware-compatible operations, enabling efficient cross-platform deployment.

Findings

Achieves speedups of 1.1x to 3.83x on X86 servers

Outperforms existing solutions like scikit-learn and pandas

Supports diverse hardware including GPU, ARM, RISC-V, and IoT devices

Abstract

The rapid development of domain-specific frameworks has presented us with a significant challenge: The current approach of implementing solutions on a case-by-case basis incurs a theoretical complexity of O(M*N), thereby increasing the cost of porting applications to different hardware platforms. To address these challenges, we propose a systematic methodology that effectively bridges the gap between domain-specific frameworks and multiple hardware devices, reducing porting complexity to O(M+N). The approach utilizes multi-layer abstractions. Different domain-specific abstractions are employed to represent applications from various domains. These abstractions are then transformed into a unified abstraction, which is subsequently translated into combinations of primitive operators. Finally, these operators are mapped to multiple hardware platforms. The implemented unified framework supports deep learning, classical machine learning, and data analysis across X86, ARM, RISC-V, IoT devices, and GPU. It outperforms existing solutions like scikit-learn, hummingbird, Spark, and pandas, achieving impressive speedups: 1.1x to 3.83x on X86 servers, 1.06x to 4.33x on ARM IoT devices, 1.25x to 3.72x on RISC-V IoT devices, and 1.93x on GPU. The source code is available at https://github.com/BenchCouncil/bridger.git.