Skip to content
High-fidelity multiphase flow simulation
Fortran Python Other
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Multi-component Flow Code (MFC)

Welcome to the MFC! The MFC is a fully-documented parallel simulation software for multi-component, multi-phase, and bubbly flows.


MFC was developed at Caltech by a group of post-doctoral scientists and graduate research students under the supervision of Professor Tim Colonius. These contributors include:

  • Dr. Spencer Bryngelson

  • Dr. Kevin Schmidmayer

  • Dr. Vedran Coralic

  • Dr. Jomela Meng

  • Dr. Kazuki Maeda

    and their contact information is located in the AUTHORS file in the source code.


The following codes are documented, please follow the links to see their Doxygen:

User's guide

A user's guide is included here.

MFC paper

The paper that describes the MFC's capabilities is located here.

Related publications

Several publications have used the MFC in various stages of its development. A partial list is included here.

Refereed journal publications:

Ph.D. Disserations:


The documents that describe how to configure and install the MFC are located in the source code as CONFIGURE and INSTALL. They are also described here.

Step 1: Configure and ensure dependencies can be located

Main dependencies: MPI and Python

If you do not have Python, it can be installed via Homebrew on OSX as:
brew install python

or compiled via your favorite package manager on Unix systems.

An MPI fortran compiler is required for all systems. If you do not have one, Homebrew can take care of this on OSX:
brew install open-mpi

or compiled via another package manager on Unix systems.

Simulation code dependency: FFTW

If you already have FFTW compiled:

  • Specify the location of your FFTW library and include files in Makefile.user (fftw_lib_dir and fftw_include_dir)

If you do not have FFTW compiler, the library and installer are included in this package. Just:
cd installers

Post process code dependency: Silo/HDF5

Post-processing of parallel data files is not required, but can indeed be handled with the MFC. For this, HDF5 and Silo must be installed

On OSX, a custom Homebrew tap for Silo is included in the installers directory. You can use it via
cd installers
brew install silo.rb

This will install silo and its dependences (including HDF5) in their usual locations (/usr/local/lib and /usr/local/include)

On Unix systems, you can install via a package manager or from source. On CentOS (also Windows 7), HDF5 binaries can be found here.

Untar this archive in your intended location via
tar -zxf [your HDF5 archive]

Silo should be downloaded here, then
tar -zxf [your Silo archive]
cd [your Silo archive]
./configure --prefix=[target installation directory] --enable-pythonmodule --enable-optimization --disable-hzip --disable-fpzip FC=mpif90 F77=mpif77 -with-hdf5=[your hdf5 directory]/include,/[your hdf5 directory]/lib --disable-silex
make install

Add the following line to your ~/.bash_profile:
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HOME/[your silo directory]/lib:/[your hdf5 directory]/lib

source ~/.bash_profile

You will then need to modify silo_lib_dir and silo_include_dir in Makefile.user to point to your silo directory.

Step 2: Build and test

Once all dependencies have been installed, the MFC can be built via

from the MFC directory. This will build all MFC components. Individual components can be built via
make [component]

where [component] is one of pre_process, simulation, or post_process.

Once this is completed, you can ensure that the software is working as intended by
make test


The MFC can be run by changing into a case directory and executing the appropriate Python input file. Example Python input files can be found in the example_cases directories, and they are called Their contents, and a guide to filling them out, are documented in the user manual. A commented, tutorial script can also be found in example_cases/3d_sphbubcollapse. MFC can be executed as
python pre_process

which will generate the restart and grid files that will be read by the simulation code. Then
python simulation

will execute the flow solver. The last (optional) step is to post treat the data files and output HDF5 databases for the flow variables via
python post_process

Note that the post-processing step requires installation of Silo and HDF5.


MFC is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License, either version 3 of the License, or any later version. A copy of the GNU General Public License is included with the MFC, and is also located at

The MFC is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

Copyright 2019 California Institute of Technology


The development of the MFC was supported in part by multiple past grants from the US Office of Naval Research (ONR), the US National Institute of Health (NIH), and the US National Science Foundation (NSF), as well as current ONR grant numbers N0014-17-1-2676 and N0014-18-1-2625 and NIH grant number 2P01-DK043881. The computations presented here utilized the Extreme Science and Engineering Discovery Environment, which is supported under NSF grant number CTS120005. K.M. acknowledges support from the Funai Foundation for Information Technology via the Overseas Scholarship.

You can’t perform that action at this time.