Question-and-Answer Resource for the Building Energy Modeling Community
 | 1 | initial version |
I’m not familiar with running parametric analyses in OpenStudio or DesignBuilder, but here’s my approach for parallel processing using multiple CPU cores in a Windows/Linux environment with Python:
Parameter Assignment: Write a function to apply specific parameters or inputs to your IDF file (see the EPPY library for reference).
Temporary Workspace: Within that function, create random temporary folders to store copies of your IDF and EPW files for each run (see the tempfile module).
Parallel Execution: Implement multiprocessing to run multiple simulations in parallel, leveraging all available CPU cores (see the multiprocessing module).
Note: Be sure to clean up by deleting all temporary folders and files once the simulations have finished. In Windows, the multiprocessing module was previously limited to a maximum of 64 CPU cores. I’m not sure if this limitation still applies.
| | 2 | No.2 Revision |
I’m not familiar with running parametric analyses in OpenStudio or DesignBuilder, but here’s my approach for parallel processing using multiple CPU cores in a Windows/Linux environment with Python:
Parameter Assignment: Write a function to apply specific parameters or inputs to your IDF file file. And assign the correct Energy+.idd path. (see the EPPY library for reference).
Temporary Workspace: Within that function, create random temporary folders to store copies of your IDF and EPW files for each run (see the tempfile module).
Parallel Execution: Implement multiprocessing to run multiple simulations in parallel, leveraging all available CPU cores (see the multiprocessing module).
Note: Be sure to clean up by deleting all temporary folders and files once the simulations have finished. In Windows, the multiprocessing module was previously limited to a maximum of 64 CPU cores. I’m not sure if this limitation still applies.
