Mr. Sebastian Schwarz 0 Report post Posted July 26, 2013 (edited) Hi everybody, I got acces to our university-cluster and can run OpenFoam on it in parallel. I wrote a script that identifies a stl-file and builds an OpenFoam-run case to be copied and executed on the cluster. After the calculation my script analyses the results and plots a results.dat. My question is how can i connect CAESES with my script to run automated optimizations for my project?Can it be done by software connector ? Best regardsSebastian Edit: The script is written in Python and runs on a Linux-machine like the cluster Edited April 6, 2018 by Joerg Palluch Share this post Link to post Share on other sites
Jörg 29 Report post Posted July 26, 2013 Hi Sebastian, Yes, you can use the software connector for this purpose. The geometry is typically provided and changed by CAESES/FFW (drag & drop it into the "input geometry" field). The script should be set at the computation of the software connector.The process is described more generalized in one of the tutorials. CheersJoerg BTW: CAESES can't run in batch mode which is sometimes required on cluster systems (FRIENDSHIP-Framework is able to run in batch mode). Share this post Link to post Share on other sites
Mr. Sebastian Schwarz 0 Report post Posted July 29, 2013 Hi Joerg, first of all "Thank You" for your fast reply. I'm sorry that i answer it now but the weekend was too hot to work. I hope I understand you right. I'm running Friendship with an Education License on my local machine. I build an local setup with the software-connector to my script. After that i can run a NelderMeadSimplex optimization for my shiphull and it finds itself the computation setup with my script? Best regardsSebastian Share this post Link to post Share on other sites
Jörg 29 Report post Posted July 29, 2013 Hi Sebastian, As far as I remember the script is searched in the project directory if it is a relative path information (mostly, we use absolute paths to the executable path). We had situations where we had to copy the script to the cluster system along with all other files. In this case you can add the script as an input file that gets copied as well. The optimization engines evaluate parameters and these parameters then refer to the computation (e.g. checking for a resistance value etc). This is the connection from the engine to the computation. In the tutorials it is also described how to extract values from a result file and how to put it into a parameter. RegardsJoerg Share this post Link to post Share on other sites