这几天,在HPC上安装调试Linux版的STAR-CCM+ 11.06.010-R8, 以串行模式加载算例时,能够成功计算。但是以并行方式加载算例时,会报出一大堆错误信息:
Design STAR-CCM+ simulation completed
Server process ended unexpectedly (return code 1)
starccm+: Rank 0:1: MPI_Init: multiple pkey found in partition key table, please choose one via MPI_IB_PKEY
starccm+: Rank 0:1: MPI_Init: pkey table:
starccm+: Rank 0:1: MPI_Init: 0x8001
starccm+: Rank 0:1: MPI_Init: 0x7fff
starccm+: Rank 0:1: MPI_Init: 0xffff
starccm+: Rank 0:1: MPI_Init: ibv_get_pkey() failed
starccm+: Rank 0:1: MPI_Init: Can't initialize RDMA device
starccm+: Rank 0:1: MPI_Init: Internal Error: Cannot initialize RDMA protocol
starccm+: Rank 0:0: MPI_Init: multiple pkey found in partition key table, please choose one via MPI_IB_PKEY
starccm+: Rank 0:0: MPI_Init: pkey table:
starccm+: Rank 0:0: MPI_Init: 0x8001
starccm+: Rank 0:0: MPI_Init: 0x7fff
starccm+: Rank 0:0: MPI_Init: 0xffff
starccm+: Rank 0:0: MPI_Init: ibv_get_pkey() failed
starccm+: Rank 0:0: MPI_Init: Can't initialize RDMA device
starccm+: Rank 0:0: MPI_Init: Internal Error: Cannot initialize RDMA protocol
MPI Application rank 1 exited before MPI_Init() with status 1
mpirun: Broken pipe
紧紧围绕着“”multiple pkey found in partition key table, please choose one via MPI_IB_PKEY“”这句话,通过"http://guge.firstguo.com/"一路来到Mellanox的社区论坛:
https://community.mellanox.com/docs/DOC-2232
里面有这么一句话:For example, the default PKEY will either have a value of 0x7FFF (limited) or 0xFFFF (full).
果断地在用户环境变量文件.bashrc中添加环境变量:export MPI_IB_PKEY=0x7FFF
再次启动starccm+, 以并行方式加载算例,成功加载,然后顺利开始并行计算!!