To interpret results, see the reference article.
» madmpi: MadMPI default configuration, on InfiniBand
» madmpi1DED: MadMPI with 1 dedicated core, on InfiniBand
» mpc: MPC, on InfiniBand
» mpc1DED: MPC with 1 dedicated core, on InfiniBand
» mpcASYNC: MPC with asynchronous progression, on InfiniBand
» mpich: MPICH 3.3, default configuration, using InfiniBand network
» mpich1DED: MPICH 3.3 with MPICH_ASYNC_PROGRESS=1 and application uses n-1 cores, using InfiniBand network
» mpichASYNC: MPICH 3.3 with MPICH_ASYNC_PROGRESS=1, using InfiniBand network
» mvapich: MVAPICH 3.2, using InfiniBand network
» openmpi: OpenMPI 4.0, on InfiniBand, default configuration
» openmpi-bxi: Bull OpenMPI 2.0, on BXI network
» openmpi-omnipath: OpenMPI 4.0, using Intel Omni-Path HFI Silicon 100 Series
» openmpiTCP: OpenMPI 4.0 on TCP
Run the benchmark on your machine, and send us your own results!alexandre.denis@inria.fr florian.reynier@inria.fr