I am a Research Assistant Professor in the College of Computing at Illinois Tech, where I work closely with Prof. Xian-He Sun at Gnosis Research Center. My research lies at the intersection of computer architecture and systems, with a focus on data-centric system design for AI and high-performance computing (HPC). I develop analytical models to characterize data-access and data-movement costs, translate these insights into memory-system and accelerator optimizations, and design memory-centric architectures and execution mechanisms for efficient computing.

Email: xlu40@illinoistech.edu

Click here to see the up-to-date version of my CV.

Research Interests

• Performance and I/O Modeling: Modeling data-movement requirements and memory-performance costs for AI/HPC workloads.
• Concurrency-Aware Memory-System Optimization: Exploiting data locality and memory-access concurrency to improve cache management, prefetching, and memory-hierarchy performance.
• Memory-Centric Architectures: Processing-in-memory (PIM), near-data processing, and heterogeneous execution for AI/HPC applications.
• Learning-Assisted System Optimization: Applying reinforcement learning and lightweight machine-learning techniques to adaptive architecture and system optimization.

Updates

• Oct 2026: I serve as a PC member for DAC 2027.
• Sep 2026: I serve as a PC member for EuroSys 2027.
• Sep 2026: Our work DynaNDE: Dynamic Near-Data Expert Scheduling for Batched MoE Inference is now available on arXiv.
• Aug 2026: Our work VIPER: Architecture-Aware Performance Modeling for Processing-in-Memory Design-Space Exploration is now available on arXiv.
• Aug 2026: Our paper Improving Data Reuse across Blocks for Efficient Block-Sparse Transformers on GPUs was accepted to PACT 2026! Congratulations to all collaborators!
• Jul 2026: Our paper HyPIM: Accelerating Hyperbolic Machine Learning via Processing-In-Memory was accepted by CCMCC 2026! Congrats to all collaborators!
• May 2026: I serve as a PC member for HPCA 2027.
• May 2026: I was invited to deliver a talk at ESSA 2026.
• Dec 2025: Our paper I/O-Aware PIM Acceleration for Long-Sequence LLM Inference with Hybrid Sparse Attention was accepted by IPDPS 2026! Congrats to all collaborators!
• Nov 2025: Our paper I/O Analysis is All You Need: An I/O Analysis for Long-Sequence Attention was accepted by ASPLOS 2026! Congrats to all collaborators! (Acceptance rate: 10.6%)
• Nov 2025: I serve as a ERC member for MLSys 2026.
• Nov 2025: Our paper Zion: A Comprehensive, Adaptive, and Lightweight Hardware Prefetcher was accepted by DATE 2026! Congrats to all collaborators!

Recent Visit