Project


  • LRBench: A semi-automatic learning rate tuning tool to improve the deep neural network training efficiency (On-going).

  • GTDLBench: A Performance Benchmark of Deep Learning Frameworks: Measure the performance of mainstream machine learning frameworks (On-going).

  • Comanche: Accelerating Deep Learning with Direct-to-GPU Storage: Integrated the Direct-to-GPU storage system into Caffe (2018/2019 Summer Intern at IBM).

  • DeepEyes: A Deep Learning Powered Localization System with Multi-modal Sensors (On-going).

  • CCAligner: A token-based code clone detector (ICSE'18).

  • LTS and Game Model Checker for PRISM: A widely applied model checker for analysis of systems.

  • C1 compiler: C1 is a subset of C and its compiler is based on LLVM.

  • WRF Optimization: The Weather Research and Forecasting(WRF) is a next-generation mesoscale numerical weather prediction system designed for both atmospheric research and operational forecasting needs. Several strategies are applied in compiling and load balance to achieve a better performance.

  • Gromacs Optimization: Gromacs is a versatile package to perform molecular dynamics. Main optimization strategies include SIMD, GPU acceleration and InfiniBand.

  • Baymini Robot: Baymini is a desktop entertainment robot able to play Go with people interactively.

  • Port Inferno to ARM Development Boards: Inferno OS was ported to a Raspberry Pi board and a Samsung S3C6410X development board.

  • A Small Web Search Engine: NLPLIR(in Chinese) and Lucene are primarily used to implement this small web search engine.

  • A Small MIPS CPU (16bit): This small CPU is implemented in Verilog and supports over 16 instructions and debugging.