Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1535968
| Summary: | MPI_OFFSET_KIND defined differently than in OpenMPI on 32bit architectures | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Dominik 'Rathann' Mierzejewski <dominik> |
| Component: | mpich | Assignee: | Zbigniew Jędrzejewski-Szmek <zbyszek> |
| Status: | CLOSED NOTABUG | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 28 | CC: | dakingun, zbyszek |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-04-04 08:43:09 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 179258, 485251 | ||
For now upstream worked around it by using MPI_ADDRESS_KIND which equals 4 in both OpenMPI and MPICH. This bug appears to have been reported against 'rawhide' during the Fedora 28 development cycle. Changing version to '28'. Reported upstream, too. Upstream bug was closed as "it's an opaque handle is because different MPI implementations can define it differently". I'll just go with upstream says. |
Description of problem: MPI_OFFSET_KIND=4 in OpenMPI while it equals 8 in MPICH on 32bit arches. I don't know which is correct, but this difference has caused compilation issues in cp2k-5.1 with MPICH: [...] /builddir/build/BUILD/cp2k-5.1/src/mpiwrap/message_passing.F:3588:70: type_descriptor%type_handle, ierr) 1 Error: Type mismatch in argument 'array_of_displacements' at (1); passed INTEGER(8) to INTEGER(4) [...] Version-Release number of selected component (if applicable): openmpi-2.1.1-5.fc28.i686 mpich-3.2.1-2.fc28.i686 How reproducible: Always Steps to Reproduce: 1. grep MPI_OFFSET_KIND= mpich-i386/mpif.h mpich-i386/mpiof.h openmpi-i386/mpif-config.h Actual results: mpich-i386/mpif.h: PARAMETER (MPI_OFFSET_KIND=8) mpich-i386/mpiof.h: PARAMETER (MPI_OFFSET_KIND= 8) openmpi-i386/mpif-config.h: parameter (MPI_OFFSET_KIND=4) Expected results: The value should be the same across MPI implementations. Additional info: https://lists.cp2k.org/private/cp2k-dev/2018-January/000856.html