Saved in:
Bibliographic Details
Main Authors: Bi, Jiahong, Schütze, Lars, Castrillon, Jeronimo
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.11966
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908829808066560
author Bi, Jiahong
Schütze, Lars
Castrillon, Jeronimo
author_facet Bi, Jiahong
Schütze, Lars
Castrillon, Jeronimo
contents Driven by the increasing demand for low-latency and real-time processing, machine learning applications are steadily migrating toward edge computing platforms, where Field-Programmable Gate Arrays (FPGAs) are widely adopted for their energy efficiency compared to CPUs and GPUs. To generate high-performance and low-power FPGA designs, several frameworks built upon High Level Synthesis (HLS) vendor tools have been proposed, among which MLIR-based frameworks are gaining significant traction due to their extensibility and ease of use. However, existing state-of-the-art frameworks often overlook the stringent resource constraints of edge devices. To address this limitation, we propose MING, an Multi-Level Intermediate Representation (MLIR)-based framework that abstracts and automates the HLS design process. Within this framework, we adopt a streaming architecture with carefully managed buffers, specifically designed to handle resource constraints while ensuring low-latency. In comparison with recent frameworks, our approach achieves on average 15x speedup for standard Convolutional Neural Network (CNN) kernels with up to four layers, and up to 200x for single-layer kernels. For kernels with larger input sizes, MING is capable of generating efficient designs that respect hardware resource constraints, whereas state-of-the-art frameworks struggle to meet.
format Preprint
id arxiv_https___arxiv_org_abs_2602_11966
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MING: An Automated CNN-to-Edge MLIR HLS framework
Bi, Jiahong
Schütze, Lars
Castrillon, Jeronimo
Hardware Architecture
Driven by the increasing demand for low-latency and real-time processing, machine learning applications are steadily migrating toward edge computing platforms, where Field-Programmable Gate Arrays (FPGAs) are widely adopted for their energy efficiency compared to CPUs and GPUs. To generate high-performance and low-power FPGA designs, several frameworks built upon High Level Synthesis (HLS) vendor tools have been proposed, among which MLIR-based frameworks are gaining significant traction due to their extensibility and ease of use. However, existing state-of-the-art frameworks often overlook the stringent resource constraints of edge devices. To address this limitation, we propose MING, an Multi-Level Intermediate Representation (MLIR)-based framework that abstracts and automates the HLS design process. Within this framework, we adopt a streaming architecture with carefully managed buffers, specifically designed to handle resource constraints while ensuring low-latency. In comparison with recent frameworks, our approach achieves on average 15x speedup for standard Convolutional Neural Network (CNN) kernels with up to four layers, and up to 200x for single-layer kernels. For kernels with larger input sizes, MING is capable of generating efficient designs that respect hardware resource constraints, whereas state-of-the-art frameworks struggle to meet.
title MING: An Automated CNN-to-Edge MLIR HLS framework
topic Hardware Architecture
url https://arxiv.org/abs/2602.11966