Researches
GPU Microarchitectural Performance ModelingOngoing
Developing analytical models and simulator-based approaches for GPU performance prediction at the microarchitectural level, focusing on instruction-level parallelism, memory hierarchy modeling, and bottleneck analysis.
AI InfrastructureOngoing
Exploring efficient system-level designs for large-scale AI training and inference, focusing on distributed computing frameworks, resource scheduling, and communication optimization to improve scalability and throughput.
Domain-Specific Instruction ExplorationFinished
Research on automated exploration and simulation of custom instructions for general-purpose processors, focusing on microarchitecture-aware exploration methods and ISA-agnostic simulation frameworks.