Avoid hwloc version conflicts by forcing pocl to load before mpi4py #169
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
inducer
approved these changes
Dec 6, 2020
It works for me, but I'd be happy if you could check if it works for you too. The test is just: $ ./install.sh --branch=fix-no-mem
$ source config/activate_env.sh
$ cd mirgecom/examples
$ python -m mpi4py ./wave-eager-mpi.py |
|
Yup, can confirm that this works (on koelsch). Thanks! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes illinois-ceesd/emirge#73
Fixes pocl-related crashes of the type:
at startup.
These happen only when running with mpi4py, e.g.
works fine with the "wrong" libhwloc version, but
does not.
The reason for the error is that since we are building mpi4py from source against the system MPI and hwloc, but pocl against conda's hwloc, the two hwloc versions might be different (v1 vs. v2). Since libhwloc v1/v2 internally use the same version number, only one version can be loaded at a time, which means that the first package that loads hwloc determines the hwloc version for everyone else.
This seems to be especially problematic when pocl is built against hwloc v1, but runs with hwloc v2 (e.g. on dunkel).
This PR makes sure that we are running with pocl's hwloc version.
Tested on dunkel and lassen.