Users December 2014

users@lists.einsteintoolkit.org

10 participants
16 discussions

Meeting Minutes for 2014-11-10
by Roland Haas 12 Dec '14

12 Dec '14

Present: Frank, Roland, Ian, Peter, Steve, Matt, Philipp, Erik ET release: * move leftover tickets marked for last release since none will be fixed before the relase * new tag "postrelease" for bugs to be handled just after the release, aim to fix them in the coming 1-2 months * direct link to this set of tickets can be accessed from the trac main page * intel restrict issue: don't use autoconf right now but instead adjust existing compiler version check. Frank, Peter, Steve to take care of. * names for release: Herschel, runners up: Wu (https://en.wikipedia.org/wiki/Chien-Shiung_Wu of the Wu experiment), Payne-Gaposchkin (https://en.wikipedia.org/wiki/Cecilia_Payne-Gaposchkin who proposed that stars are made mostly of hydrogen) Open tickets: * suggestion to spend time after release to triage tickets into important/not-so-important tickets Release frequency: * suggestion to try to receive feedback on what ET users are interested in and what ET users are currently using, however there is not agreement on whether this is a what we should ask * attempt more input and contributions from outside of core group of maintainers Yours, Roland

2 1

Wv reconstruction issue for PPM+Poly EoS
by Hee Il Kim 08 Dec '14

08 Dec '14

Hi, Even though it may not be popularly used, Wv-reconstruction for PPM+Polytropic EoS doesn't seem to work properly. It gives significantly different result compared to the one with velocity reconstruction. I have seen at least the followings are fine; 2nd order ENO + Poly EoS (cf. crashed with 3rd order ENO) PPM + Ideal Fluid EoS 5th order WENO + Ideal Fluid EoS (cf. No poly EoS supports in WENO) The tests have been done with Herschel and partly with Noether. Cheers, Hee Il

2 5

An ET Trac Ticket Describing How to Register on the ET Trac System
by Zach Etienne 08 Dec '14

08 Dec '14

Hi all, I want to submit a ticket, so I clicked the "Issue Tracker" link on http://einsteintoolkit.org/, and was redirected to https://trac.einsteintoolkit.org/. Then I tried to set up an account. There was no link to "Register", so I clicked "Login". No option to "Register New Account" was found there, either. Thus I would like to submit a trac ticket requesting that there be some link to instructions as to how to register on the trac system, on the trac website. Alternatively, clicking "Issue Tracker" could redirect to a page with instructions, as well as a link to https://trac.einsteintoolkit.org/ -Zach * * * Zachariah Etienne Assistant Professor of Mathematics West Virginia University

3 4

Meeting Minutes for 2014-12-08
by Roland Haas 08 Dec '14

08 Dec '14

Present: Roland, Frank, Yosef, Matt, Ian, Peter, Josh, Philipp Josh and Yosef report problems calling the toll free number from a phone. Tickets: * remove snprintf implementation in Cactus since Cactus requires (and tests for) C99 which always provides snprintf * patch to abort early in CST has become redundant due to Erik's recent efforts to build ExternalLibraries along with normal thorns * Jenkins now works again Ticket system: * desire to have screening of incoming tickets * talking points: ** The screening process will assign a priority to the ticket ** Description of priorities *** blocker - Cactus does no longer build *** critical - at least one thorn always fails to run *** major - needs to be fixed before release or documented in release notes *** minor - may or may not be fixed before release *** new - the initial priority created by trac to search for unclassified tickets ** Desire to have ticket status "Looked at"/"Triaged" instead of "Confirmed" that signifies that screening was done. ** Assign it to a person likely being able to help with it. This should be based on common sense, and assignation does not imply an obligation to provide a fix but is kindly requested to look at it. ** milestones to be used only during the release process to mark tickets we want to push until the release * Ian tasked with testing these ideas with ~10 existing tickets * Frank checked automated assignments are reasonable ** Change GetComponents from Eric Seidel ** Remove automated assignment from simfactory ** add "Server infrastructure" component to take over ET trac but contain all server issues * custom queries on trac ticket ** new section with most common predefined custom queries *** changes in the last week *** pending review ** new section with queries useful to review tickets ** new section with queries useful before the release Jenkins: * ask Steve to check on CCT Jenkins slave * downgrade Caltech Jenkins slave Yours, Roland

2 1

Einstein Toolkit Meeting Reminder
by Frank Loeffler 07 Dec '14

07 Dec '14

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/ In short: the number is (+1) 225-578-4942 or (+1) 866-573-0359 and the conference id is 118682#. Tickets that have been worked on (and are still open) include retiring snprintf (the Cactus version, #1712), avoiding a GRHydro build warning (#1713), and aborting early in case of errors within the CST (#1252), but the latter can probably be closed as won't fix with the split of most (all?) of the external libraries' configuration scripts into two: one to configure, and one to only build (thanks for that, Erik). Lots of smaller changes to GRHydro made the move from Zelmani to the main repository (#1152, closed, thanks to Roland), and Jenkins was not reachable, but this was fixed within minutes (#1714, closed, thanks to Barry). Besides new and otherwise mentioned tickets, we will again bring up the issue of the most convenient (and consistently working) venue for this phone call (Skype has been mentioned as not allowing US toll-free numbers from anywhere). Google Hangouts have been mentioned, but also some of their drawbacks. Also, we left the question open of how to deal with a lot of still open, and untouched bugs. We even have a ticket for that (#1698), and now would be a perfect time to do this (post-release). Another question that has been coming up was: would it be possible to have a weekly (?) full build of the ET - meaning including all external libraries that can be built on the slaves - even if system versions of external libraries are present, installed and in theory useful? This would give at least some testing for the shipped external libraries, without being too much of a burden to the slaves (hopefully). Frank Loeffler _______________________________________________ Users mailing list Users(a)einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users

1 0

Meeting Minutes for 2014-12-01
by Roland Haas 01 Dec '14

01 Dec '14

Present: Roland, Frank, Peter, Steve, Jonah, Zach, Matt, Erik, Josh Skype no longer allows calls to toll free US numbers it seems. Suggestion: can use google voice to call in. Suggestion to use Google hangouts. Updsides: better audio, video and screen sharing. Downside: max of 10 participants. Slow build of ExternalLibraries: * two patches adressing (sub-)issues ** early abort patch by Frank, can be used as quickfix ** two-stage build script by Erik that first detects settings, then builds. This needs review. Currently implemented in GSL. ** action item: go ahead with two stage procedure * related: add extra make variables at utilities build stage to simplify writing make rules for them (Roland) Goal for next release: * suggestion to add hydro analysis thorns used by Frank ** need cleanup * suggestion to get IllinoisGRMHD to ET incl. support for staggered vector potential ** needs proper solution at infrastructure level for staggering ** exampe parfile requested of the complexity level of a stable TOV * suggestion ot include random ID thorn that produces physically allowable ID for ADMBase and HydroBase. This is useful for regression tests. * remove dependency of TOVSolver on GRHydro if still present * McLachlan rewrite testing * Code to allow arbitraty number of timelevels for use by MoL and Adams-Bashforth type timesteppers Yours, Roland

1 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Users December 2014