Skip to content

Add tuning script to reduce the load time of models #140

Open
@ufechner7

Description

@ufechner7

When using MTK we cache the simplified model in a serialized problem file. The load time of the problem can be reduced a lot (lets say from 20s to 5s) when this process is pre-compiled.

Idea: Create a script "tune" that does the following:

  • create the required prob_xxx.bin file(s)
  • trigger the package pre-compilation
  • load the package and quit

Usage:

tune quasi-static=2,3,6 dynamic=2,3

--> create prob files for 2, 3 and 6 tether segments using the quasi-static model and prob files for 2 and three tether segments using the dynamic tether model and pre-compile the package such that all five problems can be loaded fast

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions