[Users] meeting minutes for 2025-12-11

[Roland Haas]

Present: Leo, Roland, Peter, Steve, Zach, Beyhan, Deborah, Johnny,
Keith, Lucas, Nikolai

Chair: Leo
Minutes: Roland

ET EU meeting 2025 discussion
=============================

* Beyhan reported on numbers of attending participants
* recordings and slides are available at
  https://sites.google.com/view/eetk25
* minutes of discussion will be added soon
* next US meeting will be at UIUC
* next EU meeting is TBD

ET 2026 release
===============

* Steve is chair
* Roland is future chair
* Beyhan agreed to help
* Deborah agrees to help with gallery runs
* Lucas volunteers to run gallery examples
* Johnny agrees to run gallery

* Roland mentions that BHNS gallery example should be in the making
* Roland suggests that having a BNS example that is not using LORENE
  and GRHydro would be nice
* Leo and Peter suggest that using FUKA as ID reader is not too
  expensive
* Beyhan suggests that SGRID would offer more options for initial data
** Steve and Peter suggest reaching out to Wolfgang Tichy for an
   introduction

Questions on the mailing list
=============================

Older messages since from before US Thanksgiving

* "Loss of convergence with subcycling in time"
  https://lists.einsteintoolkit.org/pipermail/users/2025-November/009815.html
  Steve is concerned about lack of convergence, Erik pointed out that
  full Christmas tree scheme must be used that does not use any
  interpolation in time at all

* "Safe check of variable Coordinates::volume_form_state"
  https://lists.einsteintoolkit.org/pipermail/users/2025-November/009821.html
  Suggest to create a ticket since this seems to be a bug. Storage
  itself can always be checked via CCTK_VarDataPtr

Tickets on list
===============

* "BHNS with Elliptica: Assertion `all(offset == other.offset)' failed"
  https://bitbucket.org/einsteintoolkit/tickets/issues/2899/ . No
  progress, Roland now has access to MN5.

* "CarpetX: interpolation with z-reflection symmetry only works for
  Psi4", https://bitbucket.org/einsteintoolkit/tickets/issues/2902 .
  Roland says this would be nice to have for upcoming release.

* "NoiseX: Improve noise generation in CarpetX",
  https://bitbucket.org/einsteintoolkit/tickets/issues/2855 . Moved to
  top since Roland added comment based on Erik's observation that
  vertex centered grid double cover the last grid point

* "CarpetX: Interpolator caches"
  https://bitbucket.org/einsteintoolkit/tickets/issues/2898 . Lucas is
  re-implementing so as to not use a cache but provide finer grained
  control to caller over stages (setup and interpolate)

* "GetComponents: add support to set push_url in git"
  https://bitbucket.org/einsteintoolkit/tickets/issues/2896 Steve will
  review

* "PUGH tries to free memory not allocated by malloc",
  https://bitbucket.org/einsteintoolkit/tickets/issues/2764 . No
  progress.

* "Improve McLachlan accuracy",
  https://bitbucket.org/einsteintoolkit/tickets/issues/963 . Peter and
  Zach are still working on this.


Any other business
==================

* Roland mentioned prior efforts of transitioning Cactus codes to
  github and need to reach out to current owner of cactuscode github
  user name

* Lucas brought up the issue that Perl is becoming less and less
  familiar to developers of Cactus code. Suggests to try and see if a
  transition to Python (3) maybe using a LLM to help it possible. Need
  to ensure that a common version of Python 3 is installed on all
  clusters.

We will meet once more this year next week (Dec 18th), then skip
Christmas day (25th) and New Year's day (1st).

Next chair: Deborah
Next minutes: Nikolai

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://pgp.mit.edu .