[mpich-discuss] Using Intel MPI Benchmarks

[Vaibhav Rekhate]

Hello there, I'm trying to do the performance analysis of my MPI cluster using Intel MPI Benchmarks. I am using mpich2-1.4.1 source code and Intel MPI Benchmarks 4.0 Version 2.
But I'm encountering some errors. I setup the environmental variables in make_mpich file in the src directory.
I extracted the archive and ran:
    make -f make_mpich

Part of the output where we see errors is attached below:
IMB_allreduce.o: In function `IMB_iallreduce_pure':
IMB_allreduce.c:(.text+0xd6): undefined reference to `MPI_Iallreduce'
IMB_allreduce.o: In function `IMB_iallreduce':
IMB_allreduce.c:(.text+0x246): undefined reference to `MPI_Iallreduce'
IMB_reduce_scatter.o: In function `IMB_ireduce_scatter_pure':
IMB_reduce_scatter.c:(.text+0x138): undefined reference to `MPI_Ireduce_scatter'
IMB_reduce_scatter.o: In function `IMB_ireduce_scatter':
IMB_reduce_scatter.c:(.text+0x2b8): undefined reference to `MPI_Ireduce_scatter'
IMB_reduce.o: In function `IMB_ireduce_pure':
IMB_reduce.c:(.text+0xe4): undefined reference to `MPI_Ireduce'
IMB_reduce.o: In function `IMB_ireduce':
IMB_reduce.c:(.text+0x264): undefined reference to `MPI_Ireduce'
IMB_bcast.o: In function `IMB_ibcast_pure':
IMB_bcast.c:(.text+0xc4): undefined reference to `MPI_Ibcast'
IMB_bcast.o: In function `IMB_ibcast':
IMB_bcast.c:(.text+0x244): undefined reference to `MPI_Ibcast'
IMB_barrier.o: In function `IMB_ibarrier_pure':
IMB_barrier.c:(.text+0x5e): undefined reference to `MPI_Ibarrier'
IMB_barrier.o: In function `IMB_ibarrier':
IMB_barrier.c:(.text+0x166): undefined reference to `MPI_Ibarrier'
IMB_allgather.o: In function `IMB_iallgather_pure':
IMB_allgather.c:(.text+0xee): undefined reference to `MPI_Iallgather'
IMB_allgather.o: In function `IMB_iallgather':
IMB_allgather.c:(.text+0x2a6): undefined reference to `MPI_Iallgather'
IMB_allgatherv.o: In function `IMB_iallgatherv_pure':
IMB_allgatherv.c:(.text+0x12a): undefined reference to `MPI_Iallgatherv'
IMB_allgatherv.o: In function `IMB_iallgatherv':
IMB_allgatherv.c:(.text+0x2ea): undefined reference to `MPI_Iallgatherv'
IMB_gather.o: In function `IMB_igather_pure':
IMB_gather.c:(.text+0x10c): undefined reference to `MPI_Igather'
IMB_gather.o: In function `IMB_igather':
IMB_gather.c:(.text+0x2e4): undefined reference to `MPI_Igather'
IMB_gatherv.o: In function `IMB_igatherv_pure':
IMB_gatherv.c:(.text+0x138): undefined reference to `MPI_Igatherv'
IMB_gatherv.o: In function `IMB_igatherv':
IMB_gatherv.c:(.text+0x308): undefined reference to `MPI_Igatherv'
IMB_scatter.o: In function `IMB_iscatter_pure':
IMB_scatter.c:(.text+0x10c): undefined reference to `MPI_Iscatter'
IMB_scatter.o: In function `IMB_iscatter':
IMB_scatter.c:(.text+0x2e4): undefined reference to `MPI_Iscatter'
IMB_scatterv.o: In function `IMB_iscatterv_pure':
IMB_scatterv.c:(.text+0x12d): undefined reference to `MPI_Iscatterv'
IMB_scatterv.o: In function `IMB_iscatterv':
IMB_scatterv.c:(.text+0x2f5): undefined reference to `MPI_Iscatterv'
IMB_alltoall.o: In function `IMB_ialltoall_pure':
IMB_alltoall.c:(.text+0xee): undefined reference to `MPI_Ialltoall'
IMB_alltoall.o: In function `IMB_ialltoall':
IMB_alltoall.c:(.text+0x2a6): undefined reference to `MPI_Ialltoall'
IMB_alltoallv.o: In function `IMB_ialltoallv_pure':
IMB_alltoallv.c:(.text+0x19b): undefined reference to `MPI_Ialltoallv'
IMB_alltoallv.o: In function `IMB_ialltoallv':
IMB_alltoallv.c:(.text+0x363): undefined reference to `MPI_Ialltoallv'
collect2: ld returned 1 exit status
make[1]: *** [NBC] Error 1
make[1]: Leaving directory `/work/imb/src'
make: *** [IMB-NBC] Error 2
Is there some step which I might have overlooked?
I need to check the performance for mpich2-1.4.1 only.
Please help.

Regards,
Vaibhav Rekhate
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20150401/fb0ce649/attachment.html>
-------------- next part --------------
_______________________________________________
discuss mailing list     discuss at mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss