[Mvapich-discuss] Error Encountered in MVAPICH 3.0 Configuration with oneAPI 2024 in Spack 0.21

Lieber, Matt lieber.31 at osu.edu
Thu Apr 25 12:33:16 EDT 2024


The MVAPICH team is pleased to announce the release of OSU
Micro-Benchmarks (OMB) 7.4.

Please note that OMB is also available through the Spack package
manager. Now the system administrators and users of OSU
Micro-Benchmarks (OMB) will be able to install these libraries on
their systems using Spack.

The new features, enhancements, and bug fixes for OSU Micro-Benchmarks
(OMB) 7.4 are listed here:

* New Features & Enhancements (since 7.3)

    - Add support for new benchmarks to measure network congestion
        * osu_bw_fan_in, osu_bw_fan_out
    - Add new collective benchmarks
        * osu_reduce_scatter_block, osu_ireduce_scatter_block
    - Add support for custom percentile values to evaluate benchmark performance
    - Add support to log validation failures to following benchmarks
        * Point-to-point benchmarks
            * osu_bibw, osu_bw, osu_mbw_mr, osu_latency, osu_multi_lat,
            * osu_latency_mp, osu_latency_mt, osu_bw_persistent,
            * osu_bibw_persistent, osu_latency_persistent
        * Blocking collective benchmarks
            * osu_allgather, osu_allgatherv, osu_alltoall, osu_allreduce,
            * osu_alltoallv, osu_alltoallw, osu_bcast, osu_barrier, osu_gather,
            * osu_gatherv, osu_reduce, osu_reduce_scatter, osu_scatter,
            * osu_scatterv, osu_reduce_scatter_block
        * Non-Blocking collective benchmarks
            * osu_iallgather, osu_iallgatherv, osu_iallreduce, osu_ialltoall,
            * osu_ialltoallv, osu_ialltoallw, osu_ibcast, osu_ibarrier,
            * osu_igather, osu_igatherv, osu_ireduce, osu_iscatter,
            * osu_iscatterv, osu_ireduce_scatter, osu_ireduce_scatter_block
        * Neighborhood collective benchmarks
            * osu_neighbor_allgather, osu_neighbor_allgatherv,
            * osu_neighbor_alltoall,osu_neighbor_alltoallv,
            * osu_neighbor_alltoallw, osu_ineighbor_allgatherv,
            * osu_ineighbor_allgatherv, osu_ineighbor_alltoall,
            * osu_ineighbor_alltoallv, osu_ineighbor_alltoallw
        * Persistent collective benchmarks
            * osu_allgather_persistent, osu_allgatherv_persistent,
            * osu_allreduce_persistent, osu_alltoall_persistent,
            * osu_alltoallv_persistent, osu_alltoallw_persistent,
            * osu_barrier_persistent, osu_bcast_persistent,
            * osu_gather_persistent, osu_gatherv_persistent,
            * osu_reduce_persistent, osu_reduce_scatter_persistent,
            * osu_scatter_persistent, osu_scatterv_persistent

* Bug Fixes

    - Fixed time measurement in osu_mbw_mr and osu_multi_lat benchmarks to
      consider execution time across all sender ranks.
    - Fixed validation failure when using managed device buffers.
        * Thanks to Van Man Nguyen @Eviden for report.
    - Fixed time measurement in osu_mbw_mr and osu_multi_lat benchmarks to
      consider execution time across all sender ranks.
    - Fixed multiple definition error when building with NCCL support.
        * Thanks to Klaus Noelp @FU Hagen for report.
    - Fixed bugs causing display of multiple/incorrect headers.
    - Fixed hang issue in benchmarks using multiple buffers when validation
      is enabled.
        * Thanks to Van Man Nguyen @Eviden for the report.
    - Fixed issue in osu_cas_latency benchmark that led to segmentation fault.
        * Thanks to Julien Duprat @Eviden for the report.
    - Fixed configure time failure while building with latest ROCm.
        * Thanks to Mahdieh Ghazimirsaeed @AMD for the report and patch.
    - Fixed device and managed memory allocation in one-sided benchmarks.
        * Thanks to Mikhail Brinskii @NVIDIA for the report.
    - Fixed bug in Reduce-Scatter benchmarks that led to data validation of
      only a part of the receive buffer.
    - Fixed NULL pointer error in device based one-sided benchmarks.
    - Fixed data validation issue with Reduce, Allreduce, Reduce_scatter
      benchmarks.
    - Updated configure help for "--enable-mpi4" to state that the MPI-4
      features are auto enabled.
        * Thanks to Antoine Morvan @Eviden for the report.
    - Fixed a bug in benchmarks that led to MPI_IN_PLACE support getting enabled
      only when validation is enabled.

For downloading OMB 7.4 and associated README instructions, please visit the following URL:

http://mvapich.cse.ohio-state.edu

All questions, feedback, bug reports, hints for performance tuning,
patches, and enhancements are welcome. Please post it to the
mvapich-discuss mailing list (mvapich-discuss at lists.osu.edu).

Thanks,

The MVAPICH Team

PS: We are also happy to inform you that the number of organizations
using MVAPICH libraries (and registered at the MVAPICH site) has
crossed 3,375 worldwide (in 91 countries). The number of downloads
from the MVAPICH site has crossed 1,780,000 (1.78 million).  The
MVAPICH team would like to thank all its users and organizations!!




More information about the Mvapich-discuss mailing list