
Workshop Speaker @ Supercomputing Asia 2025, Singapore
Introduction of Competition Achievements in HPC-AI 2024 as NTUHPC Representative
Introduction of Competition Achievements in HPC-AI 2024 as NTUHPC Representative
Introduced HPC educational landscape and student’s career roadmap to Former Deputy Prime Minister of Singapore
Responsible for running and optimizing ICON Climate Modeling and contributed to MLPerf Stable Diffusion inference results using Collective Mind framework.
Optimized multi-GPU training for the Llama-2 model and achieved a 6.93x training speedup on the Australia Gadi supercomputer and a 3.08x training speed up in Singapore ASPIRE2A supercomputer.
Thoroughly tested and optimized ICON code performance using different compilers, MPI, and profilers like IPM and Intel’s Vtune, leading to a 4% overall speed increase after parallelization with OpenMP.