open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear All,

These are my initial days with TELEMAC, I wanted to share a animation that I made with Telemac-2D and Tecplot for flow simulation in a confluence-section of Chicago River, Chicago, IL. The flow simulation (if you happen to see) is for about 10 kms of river length, simulated with a mesh of 22500 nodes and 40000 elements with time step of 0.5 seconds. The link to animation is here:

netfiles.uiuc.edu/xythoswfs/webui/_xy-41744429_2-t_UtjmisPo

For simulating a 1-day inflow hydrograph with the aforementioned mesh size it takes me about 10 hrs. My question is, has any one done some kind of scaling study with parallel version of Telecma-2D, till now I have only explored the serial version.

Best regards,
Sumit

Dear Sumit

Thank you for the feedback and your animation.

Yes TELEMAC-2D has been run numerous times in parallel on a wide range of networks. The following discussion shows a few numbers that have been obtained. In general the larger your model the more benefit you get from parallelisation.

The installation page explains how to setup TELEMAC for running in parallel.

Please let us know how you get on and which figures you obtain with a parallel setup.

Dear Weisgerber,

Thanks a lot for directing me to the relevant link, let me just ask one more question, has the same kind of benchmarking been done for Telemac-3D.

Best regards,
Sumit

yes of course ... for TELEMAC-3D, TOMAWAC, SISYPHE and we are starting with ARTEMIS (although we have some trouble with the MUMPS library).

Please note that the TELEMC system is being benchmarked and validated under a Continuous Integration System ... continuously.

Good luck,

Sébastien.