Summary: | sci-libs/hdf5[mpi] fails to configure | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Eugene Shalygin <eugene.shalygin> |
Component: | [OLD] Development | Assignee: | Gentoo Linux bug wranglers <bug-wranglers> |
Status: | RESOLVED INVALID | ||
Severity: | normal | CC: | cluster, sci |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
build log
config.log |
Description
Eugene Shalygin
2011-05-04 21:15:05 UTC
Created attachment 272131 [details]
build log
And what does config.log say ? O, nullmpi is from science overlay, not from the portage tree... Created attachment 272139 [details]
config.log
In such case it's probably INVALID, unless science herd cares. In-tree version supports sys-cluster/openmpi and sys-cluster/mpich2. checks are for libmpi and libmpich and if 'fortran', libmpio. As you've noticed, last check fails and stops the build. Nullmpi is not production ready and it doesn't support MPI-IO (doesn't even cover MPI standard AFAIK). It just won't work. |