SDN controller for acceleration of MPI reduction operation

This study aims to high-speed MPI_Allreduce leveraging SDN. MPI (Message Passing Interface), a de facto standard library for distributed and parallel computing, provides a suite of APIs for inter-process communication.

 

This study is being coducted mainly by the graduate students of the Graduate School of Information Science and Technology, Osaka University, Japan.

Research achievement(paper)

• Keichi Takahashi, Dashdavaa Khureltulga, Yasuhiro Watashiba, Yoshiyuki Kido, Susumu Date, Shinji Shimojo, “Performance Evaluation of SDN-enhanced MPI_Allreduce on a Cluster System with Fat-tree Interconnect”, The International Conference on High Performance Computing and Simulations (HPCS2014), pp. 784 – 792, Jul. 2014.

Related graduation thesis

• Keichi Takahashi, “Design and Implementation of Software Defined Networking Controller for Accelerating Collective Reduction Operation in MPI”, Bachelor thesis, School of Engineering Osaka University, Feb. 2013.