JEMRIS 2.9.1
open-source MRI simulations
No Matches

This section describes the installation procedures for building jemris from scratch. This is usually not necessary. Binary packages for Linux, Windows and Mac OS X, are available in the section Download.


JEMRIS relies on several libraries, which need to be downloaded and installed prior to its compilation:

  • CVODE variable time stepping ODE solver 5.8 (default package in ubuntu 22.04)
  • Xerces-C++ XML parser >=2.8
  • HDF5 a file format for storing and managing data >=1.8
  • CLN library for Numbers version >=1.2.X
  • GiNaC library for symbolic mathematics >=1.4.X
  • ISMRMRD ISMRM Raw Data Format >=1.4.X

Some Linux distributions have the some of the above packages prebuilt and included in their package management.


Parallelization with the MPI Library

Should you want to exploit more than one CPU core, you will need an implementation of MPI. JEMRIS has been tested with Open MPI and MPICH2. JEMRIS on Mac OS X works well with the pre-compiled Open-MPI implementation available from Homebrew.

Automated Image Reconstruction Pipeline

Automated image reconstruction of JEMRIS simulations requires a dedicated Docker container. The container provides several tools including the BART toolbox for image reconstruction. JEMRIS comunicates with the Docker container via a Python client, which is installed as Conda environment. The installation process requires Docker and Conda-Python installed:

  • Docker Containerized application service
  • Conda Python package manager

If one of them is missing, JEMRIS compilation and installation may still work but the image reconstruction features will not.

Boost C++

The boost C++ libraries are used for the evaluation of elliptical integration terms of the Biot-Savart law. JEMRIS is operational without the boost libraries while Biot-Savart-Loops behave like ideal coils.

Compiling JEMRIS

Download the JEMRIS source package from Download and unpack the tar ball. JEMRIS is compiled with cmake:

  • "# cd jemris-X.X"
  • "# mkdir build; cd build; cmake .."
  • "# make"
  • "# ctest -V"
  • "# make install"

The "cmake" command will also download the docker container and the Conda environment for image reconstruction, which may take some minutes depending on your internet connection. This happens only once or if there have been changes to the Docker container or the Conda environment.

The last step ("make install") will only work if you have write access to /usr/local/bin and /usr/local/share. The binaries jemris and if available pjemris, matlab GUI files and examples are installed.

After adding /usr/local/share/jemris/matlab to the Matlab path, you can run the JEMRIS Matlab GUIs for sequence design and simulation (see Sequence development tool: JEMRIS_seq and MRI Simulation tool: JEMRIS_sim).

You may start by loading one of the example sequences in the directory "/usr/local/share/jemris/exmaples".

-- last change 03.10.2023 | Tony Stoecker | Imprint | Data Protection --