CV & Publications

Research Interests

My PhD thesis focuses on cross-platform high-performance machine learning primitives and scalable hierarchical matrix (H-matrix) approximation algorithms for scientific computing and machine learning tasks. See my research page for more details.

Education Background

• B.S. in Mathematics, National Taiwan University, Taiwan, 2010
• M.S. in Mathematics, National Taiwan University, Taiwan, 2012
• Ph.D. in Computer Science, The University of Texas at Austin, 2018

Supervisors:

Prof. George Biros (ICES, PADAS research group)

and

Prof. Robert van de Geijn (CS, SHPC research group)

Employment Background

• [NOW] deep learning software engineer at NVIDIA, Santa Clara, USA
• Research Internship at NVIDIA, Santa Clara, USA, 2016 (GKMX: General Kernel Matrix Operations on GPUs, supervised by Dr. Cristopher Cecka and Dr. Michael Garland)
• Research Internship in TOTAL E&P, USA, 2014 (One-way seismic image with Fourier finite-difference migration using OpenACC, supervised by Dr. Maxime Hughes)
• Internship in Access Analytics International, Taiwan, 2008 (Porting BCS-EXT Library (Boeing) to NVIDIA GPU using CUDA, supervised by Dr. Dan'l Pierce)
• Research and teaching assistant in multiple classes and projects 2010~2017.

Honors

• Honors in NVIDIA CUDA programming contest in Taiwan, 2009 and 2013
• Dean Award, College of Science, National Taiwan University, 2012
• Student Paper Prize of The Mathematics Institution of Taiwan, 2012
• Dean's Excellence Award, University of Texas at Austin, 2013
• Finalists of Qualcomm Innovation Fellowship (26%), 2016
• Best student paper finalist of SC'18 (2%), 2018

Conferences & Workshops & Presentations

• SC'15, '17, and '18 at Austin, Denver, and Dallas
• IPDPS'16 and IPDPS'17 at Chicago and Orlando
• BLIS retreat'13~'17 at UT Austin
• UQ'12 at Oak Ridge, UQ'16 at Lausanne
• EASIAM'11 at Fukuoca, EASIAM'12 at Taipei
• Invited speaker of CUDA summer lecture, National Center for HPC, 2012.
• Invited speaker in Computational Scientific Software Development workshop, National Taiwan University, 2011
• International Workshop on Multi-Core Programming and Applications, 2010
• Taiwan-Japan Joint Workshop on Numerical Analysis and Scientific Computation, 2011
• HP(High Performance)-CAST, Taiwan, 2011

Projects

• HMLP (High Performance Machine Learning Primitives), cross-platform on x86-64, Arm, NVIDIA GPUs and Intel Xeon Phi
• GOFMM (Geometry-Oblivious Fast Multipole Method)
• Inv-Askit (Inverse Approximate Kernel Independent Treecode)
• GLACE (General Linear Algebra Computing Environment)
• CPU-GPU hybrid sparse linear solvers (multifrontal methods)
• Particle swarm optimization on GPU and its application
• Local and global optimization using LLVM
• Paxos protocol and Optimization

Super Computer Experience

• Stampede2 - KNL + Skx (TACC, UT Austin)
• Stampede - Xeon Phi + Sandy-Bridge (TACC, UT Austin)
• Maverick - K40 + Ivy-Bridge (TACC, UT Austin)
• IBM - Fermi + CPUs
• Cray XC30 - K40 + CPUs