[Users] Meeting minutes

[Roland Haas]

Meeting minutes:

Present were: Peter, Erik, Frank, Eloisa, Ian, ROland, Tanja, Bruno, Yosef

Carpet/Hg test runs:
* AEI (Ian) ran a production type inspiral simulations and finds speed 
differences around 3% and waveform differences around 10^-4 which is 
below finite differencing errors.
* however since vectorization does not yet work in Carpet/Hg and AEI 
usually uses it to obtain a 20% speed increase they will not use it in 
actual for-paper simulations
* GT (Roland) ran a hydro simulation which so far produces similar 
results in git and hg, but has not progressed far due to issues with 
learning how to use simfactory

Outstanding failing testsuites:
* RotatingSymmetry180/90 works again with Carpet/Hg
* QLM still fails (with MPI errors)

Simfactory/python:
* GT used simfactory/python for the first time, found it to be a large 
piece of software, would like more documentation and tutorials that are 
up-to-date
* reiterate ticket 221 https://trac.einsteintoolkit.org/ticket/221
* RIT needs possibility to update parameter files in an ongoing 
simulation (already possible by modifying the file directly)
* in summary:
** want more/corrected/updated documentation before the release (all 
contribute, please)
** maybe extract/incorporate documentation into python code to try and 
keep it up to date
** Erik will have a look at Ian's existing documentation 
http://www.aei.mpg.de/~ianhin/simfactory-userguide/
* wishlist (no particular order or priority):
** check for stalled simulations
** check when restarts fails continuously do to encountering the same 
error again and again
** copy input data files into simulation directory

CCE:
* try to simplify setup of parameters (relation between them)
* check if unigrid Carpet can be used instead of PUGH as well (Yosef 
thinks is should but is not quite sure)
* copy only required thorns from incoming to ET, leave out the ones that 
are not supported
* Caltech extraction thorn (Worldtube) still finds NaNs, hopes to fix by 
end of month

Fall meeting:
* see reminder email:
> - The fall workshop will be a "two-track" workshop, in the following
>   sense: There will be a ET-workshop, covering a selection of
>   mhd/eos/radiation transport questions. Here we need to decide on
>   which topic and someone to invite (but without financial support
>   at this point).
>   There will also be a Cactus-workshop, at the same time, the same
>   building, sharing the same room for part of the time, sharing some
>   discussions, and of course sharing the breaks.  This is for people
>   with only moderate interest in the ET workshop topics, but greater
>   interest in infrastructure development. Prime candidates for
>   topics here are improved Cactus scheduling and GPGPU support.
> 
>   I am happy to be able to say that coffee breaks are covered from LSU
>   side. Transportation, usual meals and hotels have to be covered by
>   the participants though. Information to suggested hotels ect isn't
>   available yet.

MHD:
* harmonic shift in McLachlan patch does not currently work, Ian and 
Bruno will work on this

Yours,
Roland

-- 
My email is as private as my paper mail. I therefore support encrypting
and signing email messages. Get my PGP key from http://keys.gnupg.net.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: OpenPGP digital signature
Url : http://lists.einsteintoolkit.org/pipermail/users/attachments/20110815/99a2dfc5/attachment.bin