Present: Roland, Qian, Bhavesh, Gabrielle, Samuel, Steve, Vassili, Yosef, Zach, Helvi, Antonio, Peter
website issues: * new website server up and running in a docker image, image is available on dockerhub in stevenrbrandt's account. https://github.com/stevenrbrandt/et-websites.git * https should now work on the website
gitter chat rooms: * up and running, have a handful of persons in it * will keep testing it
Working group updates (slides are on the agenda page: https://docs.einsteintoolkit.org/et-docs/meeting_agenda)
* IllinoisGRMHD: ** next call in May ** currently a bit swamped ** there are people actively using the code (Illinois. Frankfurt)
* Performance optimization ** group exists, has regular meetings on the last Friday of the month ** currently assessing goals and assigning priorities
* SphericalBSSN ** paper is accepted for publication, should be out soon ** working on logarithmic grid in "r" and document thorn ** will apply for inclusion in ET and move to standalone repo ** hydro and MHD are being worked on using same reference metric rescaling ** will need some help adding tests
* DataVault ** currently working on plotting of time series ** test installation ready at NCSA (development copy is here: https://datavault.dev.ncsa.illinois.edu/)
* Data Dependent Task Scheduler ** will present in SI2 PI meeting where Peter will present ** Sam is working on getting new SYNC and BC to work ** currently working on interpolation thorns ** ML_BSSN should work, Boundary should work ** currently working through test suite to identify failing tests ** expanded goal: making this work with Carpet rather than just PUGH
* Education and outreach ** postponed
* Cosmoparticle ** group has formed, meeting regularly once a month ** will have presentation on GRChombo today
US ET meeting * website is up now * registration should be possible by end of week * waiting on responses from invited speakers
Yours, Roland
Hi,
On Mon, Apr 23, 2018 at 09:48:06AM -0500, Roland Haas wrote:
- DataVault
** currently working on plotting of time series ** test installation ready at NCSA (development copy is here: https://datavault.dev.ncsa.illinois.edu/)
This is very cool!
Is there a version already available that has some data in it?
Where can I find information about the meta data that is collected together with the simulation data?
Frank
Hello Frank,
Is there a version already available that has some data in it?
Not yet, though you can upload data of your own (quota is minuscule though, this is just a development install after all).
Where can I find information about the meta data that is collected together with the simulation data?
The idea is to support the LIGO data format for waveform files. Documented here: https://arxiv.org/abs/1703.01076
We (the NCSA gravity group) are working on expanding and combining existing python and Mathematica (that being Ian Hinder's simulation tools) scripts into a single python script that would take a Cactus simulation and do all the steps to produce a LIGO hdf5 file with the waveform and which should work without problems on typical clusters.
For other types of files (ID data, checkpoints, stuff) it is less clear yet what type of metadata to support.
Yours, Roland
Hi,
On Tue, Apr 24, 2018 at 07:56:16AM -0500, Roland Haas wrote:
For other types of files (ID data, checkpoints, stuff) it is less clear yet what type of metadata to support.
What are the type of data you do target besides the waveform data? The latter is important without doubt, but I could imagine a repository of general simulation data could be quite interesting, especially if it does contain a little more than the wave data, i.e. matter quantities for simulations that contain those.
If this is still quite in the open (it did sound like that): what are the constraints the project has, especially concerning disk space and online bandwidth / data limits? And speaking of which: will it be possible to run self-written analysis on existing data "close to the data", i.e. without downloading it?
Frank
Hi Frank,
The goal of DataVault is to centralize, preserve and provide persistent and reliable access to the simulation data generated from Einstein Toolkit.
Hopefully this helps.
Best,
Qian
From: Frank Loeffler knarf@cct.lsu.edu Date: Tuesday, April 24, 2018 at 1:55 PM To: rhaas@ncsa.illinois.edu Cc: users@einsteintoolkit.org, Qian Zhang zqian1@illinois.edu Subject: Re: [Users] meeting minutes for 2018-04-23
Hi,
On Tue, Apr 24, 2018 at 07:56:16AM -0500, Roland Haas wrote: For other types of files (ID data, checkpoints, stuff) it is less clear yet what type of metadata to support.
What are the type of data you do target besides the waveform data? The latter is important without doubt, but I could imagine a repository of general simulation data could be quite interesting, especially if it does contain a little more than the wave data, i.e. matter quantities for simulations that contain those.
If this is still quite in the open (it did sound like that): what are the constraints the project has, especially concerning disk space and online bandwidth / data limits? And speaking of which: will it be possible to run self-written analysis on existing data "close to the data", i.e. without downloading it?
Frank
Hi Qian
On Tue, Apr 24, 2018 at 07:10:20PM +0000, Zhang, Qian wrote:
The goal of DataVault is to centralize, preserve and provide persistent and reliable access to the simulation data generated from Einstein Toolkit.
I knew that much. What I was after was a little more of the current status. :)
If this is still quite in the open (it did sound like that): what are the constraints the project has, especially concerning disk space and online bandwidth / data limits? And speaking of which: will it be possible to run self-written analysis on existing data "close to the data", i.e. without downloading it?
thanks, Frank
users@lists.einsteintoolkit.org
-
Frank Loeffler -
Roland Haas -
Roland Haas -
Zhang, Qian