[mpich-devel] MPICH 5.0.1rc1 released

Raffenetti, Ken raffenet at anl.gov
Tue Mar 31 10:49:32 CDT 2026


A new release candidate of MPICH, 5.0.1rc1, is now available for download. This release addresses several user-reported issues. For the full set of commits see https://urldefense.us/v3/__https://github.com/pmodels/mpich/compare/v5.0.0__;!!G_uCfscf7eWS!Yh4ciKVfeX7Nxjr_F7HPgGhkPMK9BPY_qzsZ89EmdEKOKo95eh7IRD47dQLlMQh-f8xPaFD8Vnmzz80$ …v5.0.1rc1/.

You can find the release on our downloads page (https://urldefense.us/v3/__https://www.mpich.org/downloads/__;!!G_uCfscf7eWS!Yh4ciKVfeX7Nxjr_F7HPgGhkPMK9BPY_qzsZ89EmdEKOKo95eh7IRD47dQLlMQh-f8xPaFD8kRC0sOk$ )

Regards,
The MPICH team

===============================================================================
                               Changes in 5.0.1
===============================================================================

# Fix bad cast in release-gather collectives that caused data loss
  issues on Big-Endian 64b arches (s390x)

# Fix issue with canceling MPI_ANY_SOURCE receive requests

# Fix configuration issue when C++ compiler does not support complex types

# Fix function signature issue in Hydra PBS support

# Fix crash in MPI_Allreduce with MPI_LOGICAL type

# Fix potential crash in multi-nic libfabric initialization

# Fix memory leaks in Level Zero and PMIx support

# Fix bug in CMA code when GPU support is enabled

# Add large count and other necessary aliases to ROMIO to avoid
  accidental profiling of internal MPI function usage

# Add missing error checks in rndv and colletive composition code

# Improve autogen.sh error message when autotools are too old
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/devel/attachments/20260331/23481ccd/attachment-0001.html>


More information about the devel mailing list