Release 0.9.1 - 2018-07-23

  NEW FEATURES

    - Based on Google TensorFlow 1.3.0

    - Build with Python 2.7, Cuda 8.0, CUDNN 5.0, gcc 4.8.5, and glibc 2.17

    - Compliant with TensorFlow 1.3.0 APIs and applications 

    - High-performance design with native InfiniBand support at the
      verbs level for gRPC Runtime (AR-gRPC) and TensorFlow
      - RDMA-based data communication
      - Adaptive communication protocols
      - Dynamic message chunking and accumulation
      - Support for RDMA device selection

    - Easily configurable for native InfiniBand and the traditional
      sockets based support (Ethernet and InfiniBand with IPoIB)

    - Tested with 
      - Mellanox InfiniBand adapters (e.g., EDR)
      - NVIDIA GPGPU K80