End-to-end Mapping in Heterogeneous Systems Using Graph Representation Learning

TL;DR

This paper introduces a unified graph representation learning framework for heterogeneous systems that automates code partitioning and core assignment, significantly improving execution speed and efficiency.

Contribution

The paper presents a novel end-to-end PGL framework that combines graph autoencoders and GNNs for code analysis and optimization in heterogeneous computing.

Findings

Achieved up to 6.42x speedup over thread-based execution.

Demonstrated 2.02x improvement over existing state-of-the-art methods.

Effectively predicts optimal code-to-core mappings in heterogeneous systems.

Abstract

To enable heterogeneous computing systems with autonomous programming and optimization capabilities, we propose a unified, end-to-end, programmable graph representation learning (PGL) framework that is capable of mining the complexity of high-level programs down to the universal intermediate representation, extracting the specific computational patterns and predicting which code segments would run best on a specific core in heterogeneous hardware platforms. The proposed framework extracts multi-fractal topological features from code graphs, utilizes graph autoencoders to learn how to partition the graph into computational kernels, and exploits graph neural networks (GNN) to predict the correct assignment to a processor type. In the evaluation, we validate the PGL framework and demonstrate a maximum speedup of 6.42x compared to the thread-based execution, and 2.02x compared to the state-of-the-art technique.