[Users] Meeting Minutes

[Roland Haas]

Present: Bruno, Eloisa, Seth, Josh, Tanja, Yosef, Erik, Frank, Peter,
Roland, Philipp, Manuele

Tickets:
* Roland gave a quick summary of the recent GRHydro commits
- most of them affect only MHD
- some remove old comments, remove void scheduling statements
- added support for CarpetEvolutionMask which when turned on allows
GRHydro to not warn about errors in regions that are not needed for the
evolution. Code is also a test baloon for Cray pointer support.
- all simfactory option lists should have support for Cray pointers enabled
- Christian Reisswig contributed a WENO5 reconstruction method to
GRHydro. Based on http://arxiv.org/abs/0704.2608 (code corresponds to
what is used in SpEC). Tested with shock tube and TOV.
* Tmunu's support_old_Tmunu_interface parameter changed its default
value. This requires changes to old parameter files using it. Users are
strongly encouraged to switch to the new interface (using the Tmunu grid
functions) which is much (factor of several) faster than the old
interface. Support for the old interface will be gradually phased out in
the next few releases.
* Erik reported on the flesh changes to automatically enable thorns that
were REQUIREd or OPTIONAL during compile time.
* LoopControl has been rewritten to take the hardware layout into
account, usese support library hwloc. Interface changes to C/C++ done,
Fortran outstanding.
* Carpet's processor decomposition behaviour changed due to bugfixes. No
change in data should occur but HDF5 and ASCII files might differ due to
different decomposition.
* MoL Euler integrator: Erik will commit unless vetoed. Review process
has petered out.

Process decomposition independent output:
* see discussion on trac
* generally considered a nice thing to have but not high on the priority
list

Code tarball:
* see discussion on trac ticket #1229
* lean towards gzip
* tags for release branches, all updated when release accepts bugfix patch
* tarball will not be in svn due to size (~200MB).

GRhydroMP presentation:
* Erik suggests using WebEx which is what PI uses internally (will try
out for one of the next ET calls)
* date: next week Jan 28th (unless cancelled)

Yours,
Roland

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 262 bytes
Desc: OpenPGP digital signature
Url : http://lists.einsteintoolkit.org/pipermail/users/attachments/20130121/f09ac72f/attachment.bin