Present were: Roland, Frank, Yosef, Eloisa, Josh, Tanja, Peter, Bruno
ExternalLibraries/MPI: * Yosef suggested a dummy MPI thorn to be created in Lovelace so that it easy to have a thorn compile with both releases
ET fall workshop: * location: RIT * time: week of Oct 22nd, please write to mailing list if this does not work for you * duration: to-be-discussed (currently: Mo-Fri) * topics: to-be-decided. Starting point for discussion: ** efficiency/scaling ** scheduler improvements, (reads/writes, task based scheduling, simplifying time/refinement level handling) ** generic elliptic solver * Frank will create a wiki page for the meeting
HDF5 in test suites: * Yosef sugests (and there was no loud objection) that having support via h5diff would be a good thing * possibly have student at LSU look into interfacing h5diff with the test suite perl scripts * Cactus compiles h5diff when it compiles HDF5, otherwise HDF5 has to provide a path to h5diff * previous discussion on HDF5 in testsuites is here: https://trac.einsteintoolkit.org/ticket/808
* CCE output file corruption after recovery: * Yosef has added more warning messages * error seems to be system dependent
Yours, Roland