Is It a Good Idea to Build an HLS Tool on Top of MLIR? Experience from Building the Dynamatic HLS Compiler

2026-03-20

AI Summary

This work systematically evaluates the suitability of MLIR as an infrastructure for high-level synthesis (HLS) compilers, addressing the expressiveness and optimization limitations of traditional LLVM in HLS contexts. Drawing on the industrial-scale refactoring of the Dynamatic HLS compiler, the study presents the first comprehensive validation of MLIR’s capabilities within a real-world HLS workflow. By leveraging MLIR’s multi-level intermediate representation framework, the authors integrate dataflow semantic modeling with hardware-oriented optimizations to construct an end-to-end compilation pipeline. The results demonstrate that MLIR substantially enhances modularity, development efficiency, and optimization potential in HLS tooling, while also exposing current gaps in scheduling algorithms and hardware modeling within the MLIR ecosystem—thereby offering clear directions for future research and development.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

When the MLIR project was first introduced, it promised to address the issues that the HLS community had with the LLVM project. But is this really the case, and is MLIR the "right"/"best" compiler infrastructure for HLS? We here share our experiences based on the development of Dynamatic (github.com/EPFL-LAP/dynamatic).

Problem

• HLS
• MLIR
• compiler infrastructure
• LLVM
• high-level synthesis

Innovation

• MLIR
• High-Level Synthesis
• Dynamatic
• Compiler Infrastructure
• Dataflow HLS