Users

users@lists.einsteintoolkit.org

3 participants
3984 discussions

Meeting Minutes 2021-05-27
by Gabella, William E 27 May '21

27 May '21

Present: Atul K, Bill G, Steve B, Yosef Z, Karima S, Peter D, Leonardo W, Zach E, Ken S, Maria BH, Miguel G, Roland H Chair: Atul K, Minutes: Bill G * Before agenda items? (like quick help questions, etc) None * SPEC benchmark contribution [https://bitbucket.org/einsteintoolkit/tickets/issues/2470 task] [RH, SB] Steve, Roland asked SPEC how things are going, they outlined what more they want from us. They want to be able adjust workload to keep a 40 core Haswell machine busy for many minutes. They could integrate Cactus fine, adjust config.h for various platforms ...they will do that. When testing is done, will start on step 2 proposal, workload is step 3. Their goal is for Serial max memory 1.8 GB and run for 8 min, for Parallel 40 threads 8 GB and 10 mins. They are asking for advice on adjusting workloads. Nothing is urgent but there are things to be done. See above ticket, #2470. * Release Planning [ZE] (Hendrik) Lorentz is the release name. Zach, It is coming along. Release Announcement draft has been shared among maintainers, https://docs.einsteintoolkit.org/et-docs/ET_2021_05_release_announcement . Simfactory should work with Python 2 or 3, compatible with Python 3 stated in the announcement. List of changes since last release should be double checked by everyone. Steve, Maintainers database exists inside the repo. Zach sees developers.txt, and scripts to turn that into other formats. Developers.py creates HTML, and Zenodo.py will also read developers.txt . Release helpers should go into this file for the current release and taken out for next release. Handling this file should be in the Release Planning, etc. See http://einsteintoolkit.org/developers.html and https://bitbucket.org/einsteintoolkit/www/src/master/developers.txt . Zach has been updating the Release Planning documentation https://docs.einsteintoolkit.org/et-docs/Release_Process . One advice is to word wrap at line 72 for the email formatting, he added an online tool to do that. Branches created tomorrow, so any minor changes have to happen today. Do we use SVN? Steve, maybe. Roland, we do not have any Subversion repos anymore, some checkout function does use them. Atul, Zenodo ticket #2524 should be addressed https://bitbucket.org/einsteintoolkit/tickets/issues/2524/zenodo-page-shows… . Zach, Zenodo may not be updated by May 31st. But the DOI from them is part of the bibtex key we ask people to cite. Nice if there, but not checked out for new release, so can be updated later. Zach, Will edit the main ET page, http://einsteintoolkit.org/index.html . Remove the ICERM announcement, keep North American Workshop, and Release Announcement in the center. If you are part of the release team, gallery example checkers included, double check your names at https://bitbucket.org/einsteintoolkit/www/src/master/developers.txt . ** Gallery Examples [ZE] *** [https://bitbucket.org/einsteintoolkit/tickets/issues/2172/test-binary-black… BBH] [SB] Steve, it runs and finished a bit early, so will run it to the end---ran out of time. Working with Kuibit to generate the plots vs Mathematica. Waveforms and trajectories look fine. *** [https://bitbucket.org/einsteintoolkit/tickets/issues/2175/test-single-stabl… TOV] [PD] Peter, re-ran it last night and still works fine. Will update web pages tomorrow/Friday. *** [https://bitbucket.org/einsteintoolkit/tickets/issues/2174/test-multi-patch-… Multipatch] [BG] Bill, Runs fine with and without Presync. Updated the Bitbucket / web pages, had to remind myself how to upload the tarball, what credentials to use (create a APP key on bitbucket but use my username not SSO email, go figure). *** [https://bitbucket.org/einsteintoolkit/tickets/issues/2176/test-binary-neutr… BNS] [AK], Atul, Done but not finished the Bitbucket updates. *** [https://bitbucket.org/einsteintoolkit/tickets/issues/2173/test-poisson-equa… Poisson equation] [MG] Miguel, finished the test and it was fine. Upload the tarballs and files, but need to update the links on the web page. Atul, We should re-run the examples again with the new Release...in June sometime. ** POWER [https://bitbucket.org/einsteintoolkit/tickets/issues/2413/include-power-cod… status] [RH] Maria, did look over the documentation for it and wrote something on Bitbucket and Roland should check it out. Roland, some items in the ticket #2413 are more ET related and not Power related---will clarify in the documentation. Maria, Power works with BBH but not so easy with BNS. Roland, you can run it with BNS but need to specify more items. Atul, add in documentation how to change the frequency with masses, and about quasi-local measure. Roland, want to avoid too much information for running ET in the Power documentation. ** [https://bitbucket.org/einsteintoolkit/tickets/issues?kind=bug&priority=majo… bugs of severity major or worse] Steve, #2533 is not fixed yet. If two ActiveThorns that is the problem, but need better messaging. https://bitbucket.org/einsteintoolkit/tickets/issues/2533/activethorns-not-… Yosef, #2534 Mol never worked well with Grid Arrays, but had a code that did work and it currently does not. Has not worked with 2020 releases, did in the 2019. Compile issues had missing header and maybe Presync problem. Steve, if we do fix it we need to add a test for this issue/problem so we do not break it again. https://bitbucket.org/einsteintoolkit/tickets/issues/2534/evolution-of-grid… ** all [https://bitbucket.org/einsteintoolkit/tickets/issues?milestone=ET_2021_05&s… open tickets] flagged for the next release #2526 is not a blocker and will not be ready by the release. https://bitbucket.org/einsteintoolkit/tickets/issues/2526/mclachlan-has-inc… #2469 Leonardo, I iwll push hard on this today. https://bitbucket.org/einsteintoolkit/tickets/issues/2469/none-of-the-thorn… #2488 Atul, has a minor tag so not critical right now. Zach, Last commits should be minor and not change functionality and should happen today. Will work with Steve and Roland on the Zenodo, etc. But please read the Release Announcement soon, want to finish it. Bill, we should mark the tickets (which are listed as Tasks and not Tickets) for the gallery examples as resolved when you are through with them. They will be re-opened and re-assigned for the next release. [https://www.einsteintoolkit.org/tools/unanswered.php unanswered question on mailing list] Seem to be handled. [https://bitbucket.org/einsteintoolkit/tickets/issues?status=open&status=new… open tickets sorted by update time] #2530, Roland has some ideas to fix it, not high priority and will not do anything before the release to avoid breaking anything. https://bitbucket.org/einsteintoolkit/tickets/issues/2530/thorn-vectors-fai… [https://bitbucket.org/einsteintoolkit/tickets/issues?status=new&status=open… tickets ready for review] Six tickets most still need reviewers. Please volunteer. #963, Peter, did look at it. It is a patch to an old version of McLachlan that is no longer current. Has to wait until after the release. https://bitbucket.org/einsteintoolkit/tickets/issues/963/improve-mclachlan-… * AOB Chair and Minute taker for next weeks: June 3, chair Steven B, minutes Miguel G June 10, chair Peter D, minutes Atul K 0123456789012345678901234567890123456789012345678901234567890123456789 -- ============================== William Gabella Research Assistant Professor Department of Physics and Astronomy Nashville, TN USA (o) 615-343-2713

1 0

Einstein Toolkit Meeting Reminder
by users＠einsteintoolkit.org 26 May '21

26 May '21

Hello, Please consider joining the weekly Einstein Toolkit phone call at 9:00 am US central time on Thursdays. For details on how to connect and what agenda items are to be discussed, use the link below. https://docs.einsteintoolkit.org/et-docs/Main_Page#Weekly_Users_Call --The Maintainers

1 0

Agenda for Thursday's Meeting
by users＠einsteintoolkit.org 24 May '21

24 May '21

Please update the Wiki with agenda items for Thursday's meeting. Thanks! https://docs.einsteintoolkit.org/et-docs/meeting_agenda --The Maintainers

1 0

Meeting Minutes May 20, 2021 ETK
by Hamilton, Maria 20 May '21

20 May '21

Present: Zach Etienne, Yosef Zlochower, Miguel Garcia, Leonardo Werneck, Ken Sible, Atul Kedia, Steve Brandt, Roland Haas, Peter Diener, Maria Hamilton Chair: Peter Diener Minute taker: Maria Hamilton * pre-Agenda : nothing * SPEC benchmark contribution: Steve no news * Workshop update: * Roland: deadline talks registration June 23 * ETK 2021 Workshop Webpage<https://einsteintoolkit.github.io/et2021uiuc/> * ETK 2021 Workshop Announcement<https://hyperspace.uni-frankfurt.de/2021/05/20/north-american-einsteintoolk…> * Scientific committee - Yosef, tasked to contact speakers, Roland will provide email addresses. * New release main features: * POWER: Roland just pushed a new commit, added documentation, Maria will test. * New release issues: BBH Gallery * Steve reran BBH from Gallery, run on Deep(?) Bayou cluster, code dies with assertion failure on number of threads. Made ticket, reported problem to Erik, who said is not pssibe, his code is perfect! * Zach suggested to try running with previous release, if yes, then revert to it * Steve will try, and see if works, although expects is an issue with the cluster, not teh code * Roland explained that it might come from an misunderstanding between threads and computing nodes, a problem due to hwloc1, hwloc2 conflict. He found a workaround for QueenBee and advised Steve to check the QueenBee optionlist. * Steve mentioned that hyperthreading is disabled for most clusters, while here is enabled and threads per core = 1, thus this might be the issue. * Other gallery examples: * BNS: Atul NSTracker link provided * asks what should be updated, if results identical * Roland advises update everything * Leo: works on documentation for VolumeIntegral and will update in the ETK arrangements. * Maria asks if the way VolumeIntegral is set up to follow star center does depend on mesh refinement and if is there any need for AHFinder or PunctureTracker. Zach explains that it tracks the center of the quantity set up to track, for example the lapse of a BBH system, and it does not depend of regridding. As for AHFinder or PunctureTracker, it can use both. * Zach computed constraint viloation, and set it up to avoid interior of horizon, by specifying radius and center, thus substracting off interiors, and also excising region outside some boundary like when tracking constraint violation * Name for the release ->male -> Lorentz -> no bjection, no vote necessary * Peter spoke up, Lorentz<https://en.wikipedia.org/wiki/Hendrik_Lorentz> it is * Multipach Wave Equaton -> Bill travels, sent update to Zach, bumps regarding visualization, will update/improve web page, to make it easier for next person, will update ticket on Mon * Poisson Equation: -> Miguel updated all results, will replace on the website * Roland confirms that will be no more changes of the code after the freeze * TOV -> Peter ran it, everything looks identical, can't use presink yet, and not uploaded yet. Did not know what was used to generate the plot, will use gnuplot and add script on web page, update source tree, rerun and add updates. * Zack: friendly reminder of crunch time, everything needs to be in place for next week, update all that needs to be updated. He will work on release announcement. * Release Process<https://docs.einsteintoolkit.org/et-docs/Release_Process> and Details<https://docs.einsteintoolkit.org/et-docs/Release_Details> * Roland reminded Zach to contact those who contributed and ask if they are willing to be listed - usually should be listed as authors if they are gallery runners * Zach - will take care of Zenodo link asks if its a new one or updated. Roland explains that it keeps track of versions, but needs token change to "become" Steve to be able to update. * Nothing else on the release * Unanswered mailing list queries: Peter - no unanswered email found * Tickets: * Roland couple of tickets, nothing to worry about * Flang compiler Error #2529<https://bitbucket.org/einsteintoolkit/tickets/issues/2529/flang-error-in-ca…> * Problem with Vectors compipation in gcc8, but we use gcc9 * Bug report about initialized variables, if you want some that don't exist. * Peter Erick found Flang issue * End of Agenda: Miguel reported about Gabriel's pull request for machine file Frontera simfactory2. Roland will look into it to be approved * May 27 meeting: Minute taker: Bill. Chair: Atul Maria _______________________ Maria C. Babiuc Hamilton, Ph.D. Professor, Department of Physics College of Science, Marshall University, 1 John Marshall Drive, Huntington, WV, 25755 Room S 257, Phone: (304)696-2754

1 0

Einstein Toolkit Meeting Reminder
by users＠einsteintoolkit.org 19 May '21

19 May '21

1 0

Compiling on DiRAC DIaL
by Elley, Matthew 19 May '21

19 May '21

Hi all, I am trying to compile the toolkit on DiRAC DIaL for which there is not currently a configuration script. I am not having issues with any external libraries except for PAPI, which is not present on the cluster. I've tried compiling with the ET PAPI bundle, but I get the following error: Cactus/arrangements/ExternalLibraries/PAPI/src/stats.c(20): catastrophic error: cannot open source file "papi.h" #include <papi.h> I cannot find the header file anywhere (i.e. in ET or /usr/include) so I am not sure how to proceed. Is PAPI a necessary library for ET function? Compiling with it excluded from the thornlist gives me /usr/bin/ld: cannot find -lsz make[1]: *** [/lustre/dirac/home/dc-elle1/ET/Cactus/exe/cactus_ET_dirac] Error 1 which I am not sure is connected as I am compiling from scratch. Thanks, Matt

2 1

Agenda for Thursday's Meeting
by users＠einsteintoolkit.org 17 May '21

17 May '21

Please update the Wiki with agenda items for Thursday's meeting. Thanks! https://docs.einsteintoolkit.org/et-docs/meeting_agenda --The Maintainers

1 0

Re: [Users] Query regarding visualisation software/packages
by Gabriele Bozzola 17 May '21

17 May '21

Dear Annamalai, I don't know much about yt, so the help I can offer is limited and I can only point out what you need to be able to do with yt to make a video like the one I made. I personally think that yt is not the correct tool to use for this kind of plots because you need to be able to draw several kinds of structures (curves, arrows, meshes, volume rendering), and I am not sure what is the level of support for all these different kinds of plots in yt. I used mayavi for the example. 1. Horizons are meshes (typically, 6 partially overlapping meshes). So, you need to figure out how to draw meshes with yt. You can learn about how these are represented by kuibit in the tutorial on horizons. (https://sbozzolo.github.io/kuibit/tutorials/cactus_horizons.html). I report here one of the figures to give an idea of how horizons are represented . [image: image.png] You can obtain the 3D position of all the vertices with the function `shape_at_iteration` or `shape_at_time`. 2. Spins are vectors. You have to find out how to draw arrows in yt. You can use the function `ah_origin_at_iteration` to find the center of the horizon at any given time, and then you can find the value of the vector you are interested in simply getting the quantity with brackets, e.g. h["momentum_x"] would give you a timeseries with the value of the linear momentum along the x direction for horizon h. h["momentum_x"](time) is the value of that component at the given time, which you can use to draw the arrow. 3. Orbits are 3D curves. h.ah.centroid_x is a timeseries with the x coordinate of the horizon h. You can use this to draw a 3D plot with the orbits. Read tutorial on timeseries for more information (https://sbozzolo.github.io/kuibit/tutorials/timeseries.html) 4. The curvature is the only quantity for which I think it makes sense to plot with yt. It is a 3D grid function. Normally, this has all the information about the refinement levels, but you can resample everything to a uniform grid. This can be passed to yt as "generic data", as shown here https://yt-project.org/doc/examining/generic_array_data.html. For information about grid data, have a look at the relevant tutorial https://sbozzolo.github.io/kuibit/tutorials/grid_data.html. When you use mayavi, this can be a 3D contour plot, but it still has to be resampled onto an uniform grid. 5. Gravitational waves are 2D curves that live on a different space compared to the main plot. I used matplotlib to draw this (plt.plot(strain)) and then joined the images with Pillow. The tutorial on gravitational waves will tell you more about this: https://sbozzolo.github.io/kuibit/tutorials/gravitational_waves.html 6. If you want to have something nice and user-friendly, you can take inspiration from the examples in kuibit. These examples are command-line utilities that work out-of-the-box for all the simulations. They have a lot of options and other nice features (e.g. tab-completion). Some documentation about that is here https://sbozzolo.github.io/kuibit/argparse_helper.html, but the examples themselves are great starting points. My personal recommendation for how you should move forward is to familiarize yourself with kuibit first. In this email I sketched how you can achieve what you want, but I neglected a ton of details. I suspect that it is only after you learned the basics of kuibit that this email will make sense. Also, you should ask yourself if you really want to use yt for this specific use case. If you don't want to, everything I described in this email is already implemented (with mayavi) and you can simply customize the example that I pointed out in the previous email for your needs. To get started with kuibit, you can read the tutorials, https://sbozzolo.github.io/kuibit/#tutorials and this page https://sbozzolo.github.io/kuibit/simdir.html. Then, I'd suggest trying to re-implement one of the examples on your own, for example plot_ah_separation.py ( https://sbozzolo.github.io/kuibit/examples/bins/plot_ah_separation.html). With this, you will gain enough experience to move to the next step and make the movie you want to. Every single function in kuibit is documented. In addition to the examples and tutorials, we have two additional types of documentation (for a total of four different types): - Explanations of the various modules ( https://sbozzolo.github.io/kuibit/#usage), where you can see how the modules should be used and a broad overview on how they are structured and what features they provide. - Reference and technical details of the APIs ( https://sbozzolo.github.io/kuibit/#reference-material-classes-functions), where you can see the details of the function signatures, the types expected, et cetera. You will likely need the latter two forms of documentation when you will start writing code with kuibit. Best, Gabriele On Mon, May 17, 2021 at 11:05 AM Annnamalai P S <annamalaips.97(a)gmail.com> wrote: > Dear Gabriele, > > Thank you for your reply and for taking the time to point out the > resources you felt would be relevant to my current research. > > I'm currently not very well versed with yt package but am very much > interested in the idea that you have suggested. But at the same time I'm > new to kuibit as well. If you can share any resources/documentation that > helps me to prepare data to be used for yt package that would be great. I > hope to create animated volume rendering using yt package like the ones you > have created. I also want to set it up as a user friendly and out of the > box package that can yield plots for visualization. > > On Mon, May 17, 2021 at 8:52 PM Gabriele Bozzola < > bozzola.gabriele(a)gmail.com> wrote: > >> Dear Annamalai, >> >> I am the author of kuibit, and I would like to highlight that the package >> is capable of doing what you are looking for, >> including 3D rendering. The package also comes with a lot of examples >> that you can immediately use for your >> goals. There are already some examples for 2D/3D plots of binary black >> holes, and you can see an example of >> rendering here: >> https://github.com/Sbozzolo/kuibit/tree/experimental/examples >> where I rendered a 3D movie with the apparent horizons, magnitude of the >> spins, orbits, and gravitational wave >> strain. You can produce the same movie for your simulation writing 0 >> lines of code [1]. >> >> As for yt, kuibit can prepare the data so that it is trivial to use it >> with yt (see GitHub issue >> https://github.com/Sbozzolo/kuibit/issues/18). I've never used yt and I >> don't even know where to start, but if you >> are already familiar with it, you'll be able to use it with kuibit in no >> time. [2] >> >> Finally, let me point to the documentation: >> https://sbozzolo.github.io/kuibit/. kuibit is thoroughly documented and >> there are lots of examples and tutorials that will help you get started. >> In addition to that, we have a kuibit user >> group chat on Telegram (https://t.me/kuibit) where you can ask questions >> if you find that something is not clear. >> >> Let me know if you have any questions, >> Gabriele >> >> >> [1] The code is in the experimental branch, so I would actually not >> recommend using it at the moment because it >> will break in the future. However, you can still see how it is done there >> and use the stable version of kuibit to >> achieve a similar result. At this point in time, I would recommend NOT >> using the experimental branch. >> >> [2] In that case, we would love you to contribute back to kuibit :) >> >> On Mon, May 17, 2021 at 8:05 AM Roland Haas <rhaas(a)illinois.edu> wrote: >> >>> Hello Annamalai, >>> >>> please take a look at our wiki: >>> >>> https://docs.einsteintoolkit.org/et-docs/Analysis_and_post-processing >>> >>> Linked in the "Analyzing Data Output by the Einstein Toolkit" >>> bullet of "Documentation" section of the wiki at >>> >>> https://docs.einsteintoolkit.org/et-docs/Main_Page >>> >>> Please note that yt support is not currently functional since the >>> repository linked to no longer exists. >>> >>> The most functionally complete packages are likely: >>> >>> * SimulationTools >>> * PostCactus >>> * Kuibit >>> >>> For 3d visualization I think you are pretty much limited to VisIt >>> unless you convert the HDF5 files into a format understood by other >>> packages. >>> >>> Yours, >>> Roland >>> >>> > Hi Einstein Toolkit users, >>> > >>> > I'm a relatively new Einstein Toolkit user learning to simulate and >>> > visualise binary black hole system mergers. I'm trying to write a code >>> to >>> > visualise animated trajectories, horizons, gravitational waveforms and >>> > curvature of binary system merger using Python matplotlib and >>> yt-project >>> > packages. >>> > >>> > Before I proceed, I want to know what are the list of available >>> > open-source/commercial visualisation software/packages that can produce >>> > animated plots(2D and 3D) of binary system merger with data from >>> Einstein >>> > toolkit output, and get an idea of whether something similar already >>> exists. >>> > >>> > >>> > -- >>> > Best regards, >>> > Annamalai P S >>> >>> >>> -- >>> 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 . >>> _______________________________________________ >>> Users mailing list >>> Users(a)einsteintoolkit.org >>> http://lists.einsteintoolkit.org/mailman/listinfo/users >>> >> > > -- > Annamalai P S >

2 1

Query regarding visualisation software/packages
by Annnamalai P S 17 May '21

17 May '21

Hi Einstein Toolkit users, I'm a relatively new Einstein Toolkit user learning to simulate and visualise binary black hole system mergers. I'm trying to write a code to visualise animated trajectories, horizons, gravitational waveforms and curvature of binary system merger using Python matplotlib and yt-project packages. Before I proceed, I want to know what are the list of available open-source/commercial visualisation software/packages that can produce animated plots(2D and 3D) of binary system merger with data from Einstein toolkit output, and get an idea of whether something similar already exists. -- Best regards, Annamalai P S

4 3

HelpEinsteinToolkit: "required key: user"
by jumipas＠alumni.uv.es 17 May '21

17 May '21

Dear users of Einstein Toolkit, I'm a pregraduated student in University of Valencia (Spain) and I'm using Einstein Toolkit for the first time (I used the stand-alone version of RNS inside ET but I had not big problems with it) as part of a schollarship I got this year. When I tried to run a simulation, ET returned an error which said that "consts.h" (located in ET/Cactus/arrangements/EinsteinInitialData/Hydro_RNSID/src/include) was missing. However, this file actually exists and it is located in that same directory. The error was something like this: rnsid_util.o fatal error #include "consts.h" The file consts.h does not exist or cannot be read. It was not exactly like this message, but it was quite similar. After many tries without solving this problem, I tried reinstalling the Einstein Toolkit (now I have 2) and it did not appear again, so it might be an installation error or something similar. However, now ET returns another error (actually, the machine in which I'm working is the one which returns the error, but it is from something missing in ET, not the machine). The error it returns is: Error: machine [name] is missing a required key: user Aborting Simfactory. And I have tried almost everything I know to solve it. I also spoke to some colleagues here that also use ET but all the files we compared were identical (except for the initial path to Einstein Toolkit and the username). We thought it could have something to do precisely with the username, but we changed it several times and we could not fix the error. As I still have both versions of ET (one with each error), any help with each of the errors will be much appreciated. Thank you in advance. Sincerelly, -J.

2 3

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Users