Bangtian Liu

Bangtian.jpg

Contact

Mail: liubangtian@gmail.com

Room: s-814, Meng MinWei Science Building Center for Earth System Science, Tsinghua University, Beijing, 100084

Current State

Master Candidate of Computer Science & Technology, Tsinghua University

Experience

2009 - 2013, Huazhong University of Science and Technology,CS BEng.

2013 - now, Tsinghua University, CS Master.

Research

(1) Parallel computing and Performance optimization based on heterogeneous platforms such as CPU, GPU, Xeon Phi and FPGA.

(2) High Performance Computation for geo-physics applications.

(3) Computer Arithmetic on FPGAs.


Research Experience

Optimization of Ray Tracing in Beam Migration on hybrid CPU/GPU platform. The ray tracing process have two parts: Step tracer and Observer. Step tracer is to advance the simulation of all rays by one step. Observer is to observe signaI. In my work, I optimized the step tracer on hybrid CPU/GPU platform. In the current design, we achieve 10x speedup on one Nvidia Tesla K20c GPU over 16 Intel Xeon E5-2680 CPU cores.

We develop a user-controlled FPGA design generator that can produce optimized reverse converter designs for a number of different moduli sets. Compared with the existing optimized converter designs, our proposed methods can further reduce the latency and resource consumption by 54.2% to 84.6% and 65% to 88.5% respectively.

publication

1. Bangtian Liu, Haohuan Fu, Lin Gan, Wenlai Zhao, and Guangwen Yang, Optimizing Residue Number Reverse Converters through Bitwise Arithmetic on FPGAs. The 23rd IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM 2015) Paper.

2. Conghui He, Haohuan Fu, Bangtian Liu, Huabin Ruan, and Guangwen Yang, Hui Yang, Are Osen, A GPU-based Parallel Beam Migration Design. Expanded Abstract. 2015 SEG Annual Meeting. Society of Exploration Geophysicists (SEG 2015) Paper.