Munkyu Lee Ph.D. Researcher in System Software

Introduction

Hello and welcome to my website. I’m Munkyu Lee, a researcher specializing in system software, operating systems, and GPU virtualization technologies. Currently, I am a Ph.D. researcher in the Department of Intelligent Semiconductor Engineering at Chung-Ang University, working under the supervision of Professor Cheol-Ho Hong in the System Software Lab. My research is grounded in both academic depth and practical system-level implementation, aiming to improve the performance and efficiency of large-scale computing environments.

My doctoral research focuses on spatial GPU sharing for large-scale deep neural network (DNN) inference in cloud environments. I am the first author of ParvaGPU, accepted to SC24, which proposes a novel GPU sharing framework that enhances utilization while mitigating interference across workloads. This work builds on my master’s thesis, gShare, which introduced a GPU memory sharing framework for containerized deep learning workloads.

In addition to GPU resource management, my research interests include operating system design, cloud-native infrastructure, container scheduling, and edge-cloud resource coordination. I have published multiple papers in international journals and conferences, including IEEE Transactions on Industrial Informatics, Future Generation Computer Systems, SC, and Electronics. My collaborative research has also received recognition, such as the Best Paper Award at ICCE Asia 2024.

Beyond research, I have experience in teaching and mentoring. I have served as a lecturer in Operating Systems at Sungkyunkwan University and as a teaching assistant for microprocessor circuit design and practice at Chung-Ang University. These teaching roles reflect my dedication to helping students build both theoretical knowledge and practical skills in computing.

My technical expertise includes programming in C, C++, Python, and CUDA, with experience in frameworks and platforms such as OpenCL, TensorFlow, PyTorch, Docker, Kubernetes, AWS Lambda, and NVIDIA Triton Inference Server. I have also participated in several national research projects supported by organizations such as ETRI and the Supreme Prosecutor’s Office of Korea, focusing on GPU-based container platforms and edge computing infrastructure.

This website serves as a centralized space for sharing my research outcomes, publications, and personal insights. I regularly post reflections on emerging trends in computer systems and the future of system software design. I strongly believe in collaboration and open knowledge sharing as essential drivers of innovation.

If you are interested in my work, collaborative opportunities, or would like to exchange ideas, feel free to reach out at dse112@cau.ac.kr. Thank you for visiting, and I look forward to exploring the exciting frontiers of computer science together.

Publications

Munkyu Lee, Sihoon Seong, Minki Kang, Jihyuk Lee, Gap-Joo Na, In-Geol Chun, Dimitrios Nikolopoulos and Cheol-Ho Hong, "ParvaGPU: Efficient Spatial GPU Sharing for Large-Scale DNN Inference in Cloud Environments", The International Conference for High Performance Computing, Networking, Storage, and Analysis (SC24), Nov 2024

 –  https://dl.acm.org/doi/abs/10.1109/SC41406.2024.00048

Hyunho Ahn, Munkyu Lee, Sihoon Seong, Gap-Joo Na, In-Geol Chun, Blesson Varghese, and Cheol-Ho Hong, "ScissionLite: Accelerating Distributed Deep Learning with Lightweight Data Compression for IIoT", IEEE Transactions on Industrial Informatics, Oct 2024 (JCR 2023 IF: 11.7, Top: 1.8%)

 –  https://doi.org/10.1109/TII.2024.3413340

Munkyu Lee, Hyunho Ahn, Cheol-Ho Hong, and Dimitrios S. Nikolopoulos, "gShare: A Centralized GPU Memory Management Framework to Enable GPU Memory Sharing for Containers", Future Generation Computer Systems, May 2022 (JCR 2020 IF: 7.187, Top: 6.3%)

 –  https://doi.org/10.1016/j.future.2021.12.016