present: Frank, Jonah, Roland, Josh, Ian, Erik, Christine Corbett-Moran (Caltech), Elo
Mailing list issues: * affected users were set to "bounced" status * Frank suspects this was because CCT changed firewall rules
ET release: * recursive dependency between MPI and hwloc due to routines in the thorn itself (not in the distribution tarball) * accepted solution: move respective routines into SystemStatistics thorn * add -lnuma to all linux based option lists * Erik to try and see if his testing scripts still runs, will use current einsteintoolkit.th from fro development branch * local machines/linux/osx should be tested now as well * ssl certificate of jenkis build machine is invalid, Ian will look into this (domain is owned by Barry Wardell)
Stampede: * no updates * need to decide on which compiler to use for release testing * Ian thinks that second maintenance should have brought in intel 15 * Caltech has an allocation, Christine will try and follow up locally to see if someone has already used it and could give intel 15 a try
Yours, Roland
Hello all,
Stampede:
- no updates
- need to decide on which compiler to use for release testing
- Ian thinks that second maintenance should have brought in intel 15
- Caltech has an allocation, Christine will try and follow up locally to
see if someone has already used it and could give intel 15 a try
Philipp Moesta says it compiles fine with the current option list and since the tests also passed when Erik ran them, there seems to be no pressing need to update the files this close to the release. Simfactory option lists are already considered changeable even during the lifetime of a release.
Yours, Roland
users@lists.einsteintoolkit.org
-
Roland Haas -
Roland Haas