Research
Research
My research focuses on the application of generative AI and machine learning to high-performance computing, with an emphasis on synthetic data generation, performance modeling, compiler optimization, and autotuning. I am also interested in LLM-based reasoning for GPU and systems performance.
Diffusion models and synthetic data generation for performance modeling, workload characterization, and data-efficient HPC analytics
Machine learning and generative modeling for compiler optimization selection, autotuning, and performance-aware code optimization
Evaluating large language models for architectural reasoning, GPU performance analysis, and code-level performance understanding
Publications
M. Ali and A. Qasem, “Accelerating HPC Performance Modeling via Diffusion-Based Synthetic Data Generation,” in Proc. 2026 IEEE International Conference on Cluster Computing (CLUSTER), Alexandria, VA, USA, 2026, to appear.
M. Ali and A. Qasem, “POSTER: Diffusion-Based Data Augmentation for Multi-Label Performance Modeling,” in Proc. 2026 International Conference on Parallel Architectures and Compilation Techniques (PACT ’26), Chicago, IL, USA, 2026, to appear.
M. Ali, “Robust and Portable Performance Datasets for HPC with Generative AI,” presented at the TechConnect World Innovation Conference & Expo, Austin, TX, USA, 2025.
M. Ali, Z. Sadman, and A. Qasem, “Improving Energy Efficiency of Irregular Workloads with Transformers and Tabular Data Diffusion,” in Proc. 2025 IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC), 2025.
T. Hanz, M. Ali, Z. Sadman, and A. Qasem, “Autotuning CNN Workloads on the Edge: A Hybrid Approach with Cross-Domain Embeddings,” in Proc. 2025 IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC), 2025.
M. Ali and A. Qasem, “Alleviating Dataset Constraints through Synthetic Data Generation in Machine Learning Driven Power Modeling,” in Proc. 2024 IEEE 15th International Green and Sustainable Computing Conference (IGSC), 2024. (Best Paper Finalist)
M. Ali and A. Qasem, “Towards Inductive Synthesis of Compiler Heuristics: A Case Study with Register Allocation,” presented at SC23: The International Conference for High Performance Computing, Networking, Storage and Analysis, Denver, CO, USA, 2023.
A. A. Sohan, M. Ali, F. Fairooz, A. I. Rahman, A. Chakrabarty, and M. R. Kabir, “Indoor Positioning Techniques using RSSI from Wireless Devices,” in Proc. 2019 22nd International Conference on Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 2019.