July 2024 - Nektar-users - imperial.ac.uk

Some parallel ctests hang
by Mark Dixon 14 Aug '24

14 Aug '24

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi there, We're trying to install nektar++ 5.6.0 from source on our HPC system and are having trouble passing two parallel tests: MultiRegions_Helmholtz2D_CG_P7_Modes_AllBCs_iter_ml_par3 MultiRegions_Helmholtz3D_CG_Hex_AllBCs_iter_ml_par3 In both cases they hang. Have tried openmpi, intelmpi and mvapich2. Attaching strace to the processes reveals a lot of polling and not much else. To simplify, I've reproduced this on a separate Rocky 8.10 system + EPEL packages and building against master (7b1aa23bb) in case it has already been fixed. openmpi is 4.1.1: git clone http://gitlab.nektar.info/nektar/nektar.git mkdir build cd build cmake -D NEKTAR_USE_MPI=ON ../nektar make -j24 ctest Are these tests known to be problematic, please? Thanks, Mark

4 6

Temporary restriction of new posts
by Moxey, David 31 Jul '24

31 Jul '24

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi all, Over the last few months we have been looking to work out a more sustainable solution for this mailing list, given that the mailman platform on which it is hosted is now relatively old, and will be removed from the Imperial email services in the near future. The current plan for this list is to migrate to Discourse. This is a web-based platform, but with features for still receiving notifications of posts and starting new posts via email. As well as solving the issue of maintenance, it has the significant advantage of making existing emails on this list easily searchable and accessible via the web, which is a current challenge with the mailman archives. To enable us to test this in the next couple of days, I have restricted postings on this list temporarily so that we can archive existing mail from this list and make sure it is preserved. We'll then test the new platform and can make a decision on whether to proceed. Hope to be in touch over the next few days with an update on this, but in the interim, please do make use of e.g. gitlab and email to keep in touch and ask questions. Many thanks, Dave -- Prof David Moxey, PhD MMath FHEA Professor of Computational Engineering david.moxey(a)kcl.ac.uk | web: davidmoxey.uk Department of Engineering, King's College London, Strand, London, WC2R 2LS.

1 0

Velocity Correction Scheme and Time Integration Methods
by Alexander Schukmann 12 Jul '24

12 Jul '24

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hello everyone, would somebody be so kind to briefly explain to me what exactly the TimeIntegrationScheme flag does in the session file when there is also a solver type that needs to be set? The Velocity Correction Scheme is a time integration algorithm in itself, so why exactly is setting TimeIntegrationScheme to e.g. IMEXOrder2 needed in addition to choosing VCS? All the best Alex Sicher versendet mit [Proton Mail](https://proton.me/).

2 4

finaltime and timestep
by 🌸那慢慢🌸 07 Jul '24

07 Jul '24

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hello everyone， I am a student who has just been exposed to fluid simulation and am using the base in flow-stability-cylinder to calculate the Re of vortex shedding in the flow around a square cylinder. I found that when the simulation time was extended, cases that did not originally experience vortex shedding experienced vortex shedding. I am confused why vortex shedding is related to the length of the simulation time. In addition, I understand that the finaltime should be set to 20-40 times the lift pulsation cycle, but I am not sure how to determine the timestep. I sincerely hope that you can give valuable suggestions, thank you very much! Hongtao Cheng

2 1

Adaptive SFD
by 🌸那慢慢🌸 03 Jul '24

03 Jul '24

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Dear nektar++ users and developers, I am using the Adaptive SFD method to calculate the flow field information when a square column experiences vortex shedding. But I have calculated many times according to the prompts in the user guide document, and the following two results are displayed: Vectors are lineary independent NaN found during time integration How should I set up an XML file to ensure correct calculation? I will attach my calculation file. I run the command ：IncNavierStokesSolver Cylinder-4.xml.gz Cylinder-4.xml Hongtao Cheng

4 3