Abstract:Despite the great success of high-level synthesis (HLS) tools, we observe several unresolved challenges: 1) the high-level abstraction of HLS programming styles sometimes conceals optimization opportunities; 2) the actual quality of resulting RTL designs is hard to predict; and 3) existing HLS tools do not provide flexible tradeoff (Pareto) solutions among different objectives and constraints. To this end, we propose an end-to-end framework, namely, IronMan-Pro. The primary goal is to enable a flexible and automated design space exploration (DSE), to provide either optimized solutions under user-specified constraints or Pareto tradeoffs among different objectives (such as resource types, area, and latency). IronMan-Pro consists of three components: 1) GPP, a highly accurate graph-neural-network-based performance and resource predictor; 2) RLMD, a reinforcement-learning-based multiobjective design exploration engine for optimal resource allocation strategies, aiming to provide Pareto solutions among different objectives; and 3) CT, a code transformer to assist RLMD and GPP, which extracts the data flow graphs from original HLS C/C++ and automatically generates synthesizable code with optimized HLS directives. Experimental results show that, 1) GPP achieves high prediction accuracy, reducing the prediction errors of HLS tools by in resource utilization and in critical path (CP) timing; 2) compared with meta-heuristic-based techniques, IronMan-Pro generates superior solutions improving resource utilization by and CP timing by 7.6%–16.5%; and - ) under user-specified constraints, IronMan-Pro can find satisfying solutions over 96% of the cases, more than twice as many as that of meta-heuristic-based techniques and with a speedup of up to . This work demonstrates the great potential of applying machine learning algorithms in the electronic design automation domain, especially for the hard-to-solve problems, such as timing estimation and optimization. IronMan-Pro is available at https://github.com/lydiawunan/IronMan.

Graph Neural Networks for High-Level Synthesis Design Space Exploration

Deep Inverse Design for High-Level Synthesis

AutoHLS: Learning to Accelerate Design Space Exploration for HLS Designs

HGBO-DSE: Hierarchical GNN and Bayesian Optimization Based HLS Design Space Exploration

Cross-Modality Program Representation Learning for Electronic Design Automation with High-Level Synthesis

High-level Synthesis Performance Prediction Using GNNs

IronMan-Pro: Multiobjective Design Space Exploration in HLS via Reinforcement Learning and Graph Neural Network-Based Modeling

High-Level Synthesis Performance Prediction Using GNNs: Benchmarking, Modeling, and Advancing

Learning to Compare Hardware Designs for High-Level Synthesis

DB4HLS: A Database of High-Level Synthesis Design Space Explorations

Learning from the Past: Efficient High-level Synthesis Design Space Exploration for FPGAs

Scheduling Information-Guided Efficient High-Level Synthesis Design Space Exploration

HLPerf: Demystifying the Performance of HLS-based Graph Neural Networks with Dataflow Architectures

Multi-objective Design Space Exploration for High-Level Synthesis Via Bayesian Optimization

Efficient Task Transfer for HLS DSE

Heterogeneous Graph Neural Network-Based Imitation Learning for Gate Sizing Acceleration.

A Survey of Graph Neural Networks for Electronic Design Automation

IronMan: GNN-assisted Design Space Exploration in High-Level Synthesis Via Reinforcement Learning.

GNNHLS: Evaluating Graph Neural Network Inference via High-Level Synthesis

GANDSE: Generative Adversarial Network based Design Space Exploration for Neural Network Accelerator Design

GAHLS: an optimized graph analytics based high level synthesis framework