Users June 2011

users@lists.einsteintoolkit.org

14 participants
32 discussions

Some thoughts on the webpage..
by azebrowski＠cct.lsu.edu 13 Jun '11

13 Jun '11

Hi all, I'm currently overhauling some code for a Cactus project, and thought it'd be a good idea to update my install with the newest ET release. Here are a few things I've noticed while looking for the newest release: 1) The "releases" button/page contains a lot of information on the latest release (Curie), but no actual link to either the component list itself or a page detailing download and installation directions. This seems like an oversight that would be easy to correct. 2) The "documentation" header does not contain a link to Cactus tutorials. Perhaps there could be a "Quickstart" label or a "Getting Started" label, which then loads a page that gives a choice between the beginner/advanced tutorials? A new user might click on the "User Guide" from the main menu thinking that's all there is to get started with and feel overwhelmed, when all s/he might want is to just get their hands dirty with an example or two. Anyways, hope this is useful criticism, just saw some things that I thought could be cleared up a bit :) -Ashley Zebrowski

2 1

Error from AEIThorns/ADMMass
by Scott Hawley 13 Jun '11

13 Jun '11

INFO (TexMEx): Done with TexMEx INFO (ADMMacros): Spatial finite differencing order: 4 INFO (ADMMacros): Spatial finite differencing order: 4 INFO (ADMMacros): Spatial finite differencing order: 4 INFO (ADMMacros): Spatial finite differencing order: 4 WARNING level 0 in thorn ADMMass processor 1 host hedges.belmont.edu (line 62 of /home/shawley/Cactus/configs/texmex/build/ADMMass/volume_integral.c): -> Error in obtaining OutsideMask state descriptors The relevant lines from the parameter file are SpaceMask::use_mask = yes LegoExcision::fixed_excision = "sphere" LegoExcision::num_fixed_regions = 2 LegoExcision::fixed_size = 0.9 LegoExcision::fixed_origin_x = -5.0 LegoExcision::fixed2_size = 1.0 LegoExcision::fixed2_origin_x = 5.0 ADMMass::ADMMass_volume_radius[0] = 30 ADMMass::ADMMass_Excise_Horizons = "yes" Does anyone know how to fix this? Thanks, Scott Attachments: Attached are the full parameter file and a full output log. -- Scott H. Hawley, Ph.D. Asst. Prof. of Physics Chemistry & Physics Dept Office: Hitch 100D Belmont University Tel: +1-615-460-6206 Nashville, TN 37212 USA Fax: +1-615-460-5458 PGP Key at http://sks-keyservers.net

2 3

Meeting minutes
by Roland Haas 13 Jun '11

13 Jun '11

Meeting minutes: Present were: Tanja, Roland, Erik, Frank, Josh, Christian, Yosef, Béla Szilágyi, Maria Babiuc CCE code: * thorns will (did) appear today in https://svn.einsteintoolkit.org/incoming/ arrangement suggestions are (as in in Frank's email): SphericalSlice -> CactusNumerical (analog SphericalSurface) new arrangement PITTNull (?) for CCE evolution code WorldTube -> PITTNull (CCE extraction thorn) ADMDerivatives -> EinsteinBase/EinsteinUtils (?) * please look at thorns once they appear and start discussion on mailing list * agree that PITTNull is the correct thorn name (since PITT was used in the papers) * CCE evolution is done in a Cactus thorn that uses PUGH * work cycle is to run the main simulation with an "interface" thorn that writes out initial data for the characteristic evolution, then run a postprocessing run using PUGH and the CCE thorn * two (interface) thorns to initial extract data for CCE ** original extraction code used in 1011.4223 by Yosef (stand-alone ?) *** tested with LazEv, currently being tested with McLachlan ** WorldTube by Christian Reisswig (requires extra infrastructure thorns) *** tested with MacLachlan and the AEI multipatch code (Llama?) * ask Christian Reisswig for short description of WorldTube, then decide whether to ** keep only one thorn ** keep both thorns ** merge them * requested citations for PITTNull are in README: > If you use this code, then we request that you cite the following papers: > > Title: Characteristic Extraction Tool for Gravitational Waveforms > Authors: M.C. Babiuc, B. Szilagyi, J. Winicour, Y. Zlochower > arXiv:1011.4223 > submitted to Physical Review D > > Title: Cauchy-characteristic matching > Authors: N. Bishop, R. Isaacson, R. Gomez, L. Lehner, B. Szilagyi, J. Winicour > arXiv: gr-qc/9801070 > Appeared in "On the Black Hole Trail" > eds. Bala Iyer and Biplab Bhawal (Kluwer) * license still needs to be specified in README file new infrastructure thorns: * SphericalSlice: a mini driver operations on a sphere (IO, coordinates) * ADMDerivatives: computes time and space derivatives of the ADM variables * status of examples in ET paper: ** TOV runs still in queue on QueenBee ** high res single distorted black hole has not been completed yet (there was quite a lot of waiting time on Kraken) but has made some progress (end reworded quote) * stellar collapse example: Christian will work on it fall ET workshop: * Josh will send around old proposal for workshop to be used as starting point * organizers will decide on preliminary budget and speaker list for purposes of writing the skeleton proposal Yours, Roland

1 0

Grid placement utility?
by Scott Hawley 13 Jun '11

13 Jun '11

Just curious: Has anyone developed a user-friendly utility for placing refined grids for Carpet? I'm imagining writing a little "drawing" program which will allow the user to draw rectangles to place grids, and then it would generate the CarpetRegrid::coordinates lines for a parameter file. But I don't want to reinvent the wheel if someone's already done something similar. Thanks, Scott -- Scott H. Hawley, Ph.D. Asst. Prof. of Physics Chemistry & Physics Dept Office: Hitch 100D Belmont University Tel: +1-615-460-6206 Nashville, TN 37212 USA Fax: +1-615-460-5458 PGP Key at http://sks-keyservers.net

3 2

Einstein Toolkit Meeting Reminder
by Frank Loeffler 13 Jun '11

13 Jun '11

Hi Please consider joining the weekly Einstein Toolkit phone call at 10 am US central time on Mondays. As usual, you can find instructions how to join on the following web site: http://einsteintoolkit.org/community/support/ I short: the number is (+1) 225-578-4942 or (+1) 866-573-0359 and the conference id is 118682#. Points of discussion: 1) Inclusion of CCE code into the Einstein Toolkit We planned to have the code ready in /incoming for people to look at but we'll have to do that tomorrow. This will mean to introduce new thorns, and a new arrangement: SphericalSlice -> CactusNumerical (analog SphericalSurface) new arrangement PITTNull (?) for CCE evolution code WorldTube -> PITTNull (CCE extraction thorn) ADMDerivatives -> EinsteinBase/EinsteinUtils (?) I suggest we make the code available within the coming week in incoming/ and decide about the final locations, so that we might already be able to include it next week, assuming we are able to resolve all of the remaining dependency/cleanup/documentation issues. 2) ET paper: - Mostly status on example sections. TOV: still sitting in the queue on queenbee Frank Loeffler

2 1

EinsteinBase/TmunuBase: request to make parameters restricted
by Bruno Giacomazzo 09 Jun '11

09 Jun '11

Hi, could someone apply the attached patch to EinsteinBase/TmunuBase? It just makes the parameters restricted, so that other thorns can easily use their values (e.g. Whisky checking if Tmunu has been stored in memory,...). I use it already in my version of Cactus/Whisky and I didn't encounter any problem. Thanks, Bruno Dr. Bruno Giacomazzo Department of Astronomy University of Maryland College Park MD 20742-2421 USA Tel. : +1 301 405 0676 Fax : +1 301 314 9067 Tel. : +1 301 286 3809 (NASA/GSFC) Fax : +1 301 286 7230 (NASA/GSFC) email : bgiacoma(a)astro.umd.edu web: http://www.brunogiacomazzo.org ---------------------------------------------------------------------- There are only 10 types of people in the world: Those who understand binary, and those who don't ----------------------------------------------------------------------

3 2

RotatingSymmetry* test cases breakage
by Erik Schnetter 08 Jun '11

08 Jun '11

Over night, the RotatingSymmetry* test cases broke. I believe this may be due to a recent correction in Carpet HG that corrected the setup of the reduction mask. -erik -- Erik Schnetter <schnetter(a)cct.lsu.edu> http://www.cct.lsu.edu/~eschnett/

2 1

Re: [Users] [Einstein Toolkit] #437: Correct autoconfiguration of restrict keyword; add support for builtin_expect
by Erik Schnetter 08 Jun '11

08 Jun '11

On Sun, Jun 5, 2011 at 2:59 PM, Einstein Toolkit <trac-noreply(a)einsteintoolkit.org> wrote: > This sounds like potentially quite a nice addition. How difficult would it > be to take this further and enable the use of profile guided optimization > (eg. http://stackoverflow.com/questions/2738835/learning-sample-of-likely- > and-unlikely-compiler-hints). Do you think it would be likely to help > much? Profile-guided optimisations need to be handled differently, namely via compiler options, not via source code annotations. One also needs a representative test case. I assume the main problem with this is to ensure that the compiler can make the necessary connections between executable, object files, and source files, which often fails since Cactus places object files in libraries. Apart from this, PGO would just involve experimenting with compiler options to see whether one can obtain a performance benefit. -erik -- Erik Schnetter <schnetter(a)cct.lsu.edu> http://www.cct.lsu.edu/~eschnett/

1 0

Einstein Toolkit Meeting Reminder
by Frank Loeffler 07 Jun '11

07 Jun '11

2 2

Meeting minutes
by Roland Haas 06 Jun '11

06 Jun '11

Meeting minutes: Present were: Tanja, Roland, Ian, Erik, Peter, Frank ET paper: * status of examples: ** TOV runs are finished (low and high resolution). Fundamental mode oscillations are long lived and can be observed nicely. Frequency matches expectation. Higher modes are not easy to see, have seen none yet. ** single black hole runs have finished, will process them once Kraken is online again (did I get this right?) ** BBH runs are done, L2 norm of masked constraints does not converge to zero with higher resolution. Suggest plotting 1D plots of the constraint to identify if there is a small region that spoils convergence. transition to Carpet/Mercurial: * Ian Hinder has created a wiki page where we can collect information on how to transition to the new Carpet version: https://docs.einsteintoolkit.org/et-docs/Carpet_Transition * please add any hints that you have there * Carpet/Hg changes the way buffer zones are handled, it is now the regridding thorn's (ie CarpetRegrid2's) responsibility to add sufficient buffer zones to any region that it passes to Carpet/CarpetLib. Carpet then removes a layer of zones from the input region and declares them buffer zones. * currently some testsuites (http://www.aei.mpg.de/~ianhin/tests/einsteintoolkit/) fail with the Hg version but not the git version. * some failures are obvious/already fixed (Slab,Cartoon), others are unclear (WeylScal4, however there the tests use CarpetRegrid not CarpetRegrid2 which is likely the cause) * Carpet/Hg seems to not work with multipatch anymore (failures after few iterations), Ian and Erik will get together via phone to look into this * Bruno reported (see mailing list) issues with RotatingSymmetry and Carpet/Hg ** suspect this to be related to boundary conditions not being properly applied, Carpet/Hg poisons with NaN by default, Carpet/git does not. ** Ian will re-run Bruno's troublesome BBH parameter file * will create tickets for each problem fall ET workshop: * dates: Oct 14th-18th * either at LSU or GT. LSU is known to have space, GT is checking availability of rooms * preliminary program is on wiki: https://docs.einsteintoolkit.org/et-docs/ET_Workshop_Fall_2011 Yours, Roland

2 1

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Users June 2011