Abstract:Coarse-grained reconfigurable architecture (CGRA) is an emerging computing architecture that provides a trade-off between energy efficiency and flexibility. Although the CGRA mapping problem has been explored for many years, efficiently mapping complex loop kernels to large-scale CGRAs is still challenging. In this paper, we present GRAFT, GRaph neural network (GNN) Adaptive Framework for efficienT CGRA Mapping, that can be tuned automatically to generate high-quality mapping for various CGRAs under a graph-based architecture two-level intermediate representation (IR). The low-level IR binds computing resources with routing resources annotated with architectural information, while the high-level IR provides a concise way to describe hardware modules and elastic interconnections among them. GRAFT consists of two independent components, 1) GNNbased placement distribution prediction and 2) a progressive graph-based mapper. We abstract the architecture in graph and use GNN to analyze the graph attributes and identify the subgraph relationships between data flow graphs (DFGs) and the architecture graph, providing a panoramic view for the agile mapper. In addition, we enhance the graph-based mapping algorithm that introduces a local cache and pruning strategies and takes into account the impact of DFG's placement distribution on the architecture. Experiments show that GRAFT achieves a 64.5x faster mapping speed with better performance on large-scale CGRA compared to the state-of-the-art GNN framework LISA. Moreover, GRAFT can attain 1.5x higher resource utilization with a 122x mapping speedup over Morpher by architecture tuning.

Environment ( CGRA ) Agent ( Mapping algorithm ) Action : PE Interchange Reward :

Data-Flow Graph Mapping Optimization for CGRA with Deep Reinforcement Learning

Learning to Cooperate: Application of Deep Reinforcement Learning for Online AGV Path Finding.

TransMap: an Efficient CGRA Mapping Framework Via Transformer and Deep Reinforcement Learning

Towards High-Quality CGRA Mapping with Graph Neural Networks and Reinforcement Learning

E2EMap: End-to-End Reinforcement Learning for CGRA Compilation Via Reverse Mapping

DFGNet: Mapping Dataflow Graph Onto CGRA by a Deep Learning Approach

OrderMap: A CGRA Mapping Algorithm based on DFG Analysis and Graph Homomorphism

GoodFloorplan: Graph Convolutional Network and Reinforcement Learning-Based Floorplanning

An Architecture-Agnostic Dataflow Mapping Framework on CGRA

MapZero: Mapping for Coarse-grained Reconfigurable Architectures with Reinforcement Learning and Monte-Carlo Tree Search

GRAFT: GNN-based Adaptive Framework for Efficient CGRA Mapping

Reinforcement Learning Approach for Mapping Applications to Dataflow-Based Coarse-Grained Reconfigurable Array

Routability-Enhanced Scheduling for Application Mapping on CGRAs

Action Mapping for Reinforcement Learning in Continuous Environments with Constraints

HierCGRA: A Novel Framework for Large-Scale CGRA with Hierarchical Modeling and Automated Design Space Exploration

Fast mapping algorithm with optimized placement selection for CGRA

A Novel CGRA Architecture and Mapping Algorithm for Application Acceleration

A Deep Reinforcement Learning Agent for Geometry Online Tutoring

An Efficient and Flexible Stochastic CGRA Mapping Approach

Mapping DAG to CGRA Using Min-Max Ant Colony System