PyOPUS - Simulation, Optimization, and Design

• Home
• Screenshots
• Manual
• Download
• Building
• FAQ
• About

PyOPUS 0.11.2 downloads

Sources tarball

• PyOPUS 0.11.2 sources (including documentation sources and demos)

Sources on Codeberg

• https://codeberg.org/arpadbuermen/PyOPUS/src/tag/v0.11.2

PyOPUS depends on:
Python interpreter (version 3.11) and libraries: PyQt5, pyqtgraph, lxml, NumPy, SciPy, MatPlotLib, Greenlet, CVXOPT, Pandas, and (optional) datatable.
Note that the datatable library is not available under Debian 12 (Bookworm) and must be installed via pip3 from PyPI.

See building instructions on how to build and install PyOPUS from sources.

Wheel files

• PyOPUS 0.11.2 for Linux (AMD64)
• PyOPUS 0.11.2 for Windows (AMD64)

The Python interpreter package was obtained from the official Python website.

Read section Installation for Linux in the README file on how to install the wheel file under Linux. For Windows read section Installation for Windows (from your own package sources). If you are a beginner or experince problems, we recommend that you use the Window quickstart packages (see below).

Html documentation and demos

• PyOPUS 0.11.2 html documentation and demos

Windows quickstart packages (you need this if you are using Windows and want to get started quickly)

These packages include everything (Python intepreter, libraries, and PyOPUS).

• PyOPUS 0.11.2 Windows quickstart package for AMD64

Read section Installation for Windows (quickstart) in the README file.

Configuring KiCad for use with the PyOPUS netlister

1. Select Tools/Generate Netlist File (Eeschema 5) or File/Export/Netlist (Eeschema 6).

2. Click "Add Plugin". Under "Title" enter Spice Opus.

3. Under "Netlist command" enter:

python3 -m pyopus.netlister.kicad -i "%I" -o "%O"

4. For Windows the command for running the netlister is:

C:\Python310\pythonw.exe -m pyopus.netlister.kicad -i "%I" -o "%O"

5. Click "OK". A tab will appear for generating Spice Opus netlists.

This procedure needs to be done only once.

Preparing PyOPUS for parallel processing with MPI

If you intend to run parallel processes on your local machine, there is nothing further to do.

If you intend to run parallel processes on a computer farm and your home folder /home/ is nfs-mounted on all worker computers as /home/ there is nothing further to do. If this is not the case see the documentation of pyopus.parallel.base and set the

PARALLEL_MIRRORED_STORAGE 

environmental variable on all computers participating in parallel processing.

If you want your workers to use local copies of files (mirroring) for faster processing and you intend to use /tmp folder for storing the local copies, there is nothing further to do. If you want to store the local copies somewhere else see the documentation of pyopus.parallel.base and set the

PARALLEL_LOCAL_STORAGE 

environmentral variable on all computers participating in parallel processing.

Under Windows you must start

smpd -d 3

in a command prompt on all hosts where you want to start MPI processes before starting any parallel runs. The ``-d 3`` option will produce debug output which will help you solve problems with your MPI setup. The smpd executable is a part of Microsoft MPI.

Parallel processing on a computer farm was not tested under Windows. This is uncharted territory. Reports are welcome.

Older versions of PyOPUS
Note that versions prior to 0.8 require Parallel Virtual Machine (PVM) for parallel processing.

• PyOPUS 0.6
• PyOPUS 0.7
• PyOPUS 0.8
• PyOPUS 0.9
• PyOPUS 0.10
• PyOPUS 0.11
• PyOPUS 0.11.1