• Tensor Model and Computation

    High-performance tensor theory and deep learning applications on GPU hardware.

    Read more
  • Large-scale Graph Data Processing

    Computation model and framework for incremental graph computation on GPU clusters.

    Read more
  • Powerful Many-Core GPUs

    Exploit commodity GPU hardware to accelerate data processing and artificial intelligence.

    Read more

Research Statement

Intelligent Information Processing(IIP) Lab (智能信息处理研究室) was founded by Dr. Tao Zhang in 2015.

The IIP Lab conducts research in all aspects of computer systems, with a primary focus on:

  • Big Data Processing

  • Machine Learning Algorithms and Applications

  • GPU Heterogeneous Computing

  • Intelligent Information Processing

Our research also involves other disciplines such as computer architecture and general parallel computing systems. The research themes of IIP Lab are improving the performance, scalability, productivity, availability, and reliability of distributed and centralized computer systems.

We have built a number of systems and software tools to improve the performance and engery efficiency of computer systems. Examples include a single-PC hybrid CPU and GPU graph computation platform (gGraph), a distributed graph computation platform on heterogeneous CPU and GPU cloud (HGraph), an open-source software library for characterizing and load balancing irregular applications on GPUs (cuIRRE), and two energy-efficient GPU architecture (Buddy SM) and (eDRAM-Based On-Chip storage).

Now, there are three faculty members, four master students and several undergraduate students in the lab. Our lab is supported by : Natural Science Foundation of China, Natural Science Foundation of Shanghai,Shanghai Innovation Action Plan Project, and Shanghai Educational Commission. And several collaborative research projects are going on between our lab and other research institutes including University of New Mexico, and Shanghai Jiao Tong University.

Latest Research

Tensor completion on GPUs

Accelerate low-tubal-rank tensor completion on a Nvidia M40 GPU and gain 20x and higher speedups.

Read more

Incremental graph computing on GPUs

The CPU and GPU processors work hybridly using the Update-Gather-Apply-Scatter computation model iteratively.

Read more

Join US

The lab is looking for collaborative researchers such as PhD or master students who are:
  • interested in machine learning, GPU heterogeneous computing, and Big data processing.
  • with good skills on mathematics(especially statistical math.), algorithms, and programming.
  • with good English reading and writing skills.
  • interested in research.
We provide:
  • guidances on research.
  • training on academic English paper writing.
  • experiences on research and engineering projects.
  • experiment equipments and so on.
Welcome joining! You can contact us by email(taozhang [at] shu [dot] edu [dot] cn) or pay a visit to our lab.
School of Computer Engineering and Science,Shanghai University
333 Nanchen Road, BaoShan District, Shanghai 200444, China.
© Copyright 2015-2018 IIP Lab @ SHU All rights reserved.