Solver Execution

The Solver Framework has a great demand for user supplied settings because it interfaces with a lot of external software and every user envrionment is a little bit different. A more sophisticated abbroach than just using FreeCADs Parameter System directely was taken to tackle this problem.

The run Module

Execute Solver and obtain Reports and Results.

Integral part of the Solver Framework which contains components responsible for executing the solver in the background. Also provides an asynchronous communication system with the solver running in the background. The purpose of this module is to be as generic as possible. It can execute every solver supported by the fem workbench. The threading and communication support is mainly implemented by the femsolver.task and femsolver.signal modules.

exception femsolver.run.DirectoryDoesNotExistError
exception femsolver.run.MustSaveError
femsolver.run.getMachine(solver, path=None)

Get or create Machine using caching mechanism.

Parameters

  • solver – A document object which must be a famework complient solver. This means that it should be derived from the document object provided by femsolver.solverbase and implement all required methods correctely. Of particular importance is getMachine as it is used by this method to create a new Machine on cache miss.

  • path – A valid filesystem path which shall be associetad with the machine.

femsolver.run.run_fem_solver(solver, working_dir=None)

Execute solver of the solver framwork.

Uses getMachine to obtain a Machine instance of the solver. It than executes the Machine with using the RESULTS target (see Machine for infos about different targets). This method is blocking, it waits for the solver to finished before returning. Be aware of Machine caching when using the function.

Parameters

  • solver – A document object which must be a famework complient solver. This means that it should be derived from the document object provided by femsolver.solverbase and implement all required methods correctely. Of particular importance is getMachine as it is used by this method the get the Machine used to execute the solver.

  • working_dir – If specified it overwrites the automatic and user configurable working directory management of the Solver framework. Should always be a absolute path because the location of the binary is not consistent among platforms. If None the automatic working directory management is used.

Note

There is some legacy code to execute the old Calculix solver (pre-framework) which behaives differently because it doesn’t use a Machine.