Installation Guide

Recent changes

See Recent changes for a list of recent updates to SKIRT and PTS.

Types of installation

SKIRT and the Python Toolkit for SKIRT (PTS) can be installed on any recent 64-bit version of Unix, macOS, and Microsoft Windows (for more details, see System requirements below). Your approach will depend on your usage intentions and computer system. We can distinguish two main types of environment:

• A laptop or desktop computer running macOS, a Linux distribution such as Ubuntu, or Microsoft Windows. In this environment you probably want to install both SKIRT and PTS, and you are likely to select tools that offer a graphical user interface. If you'd like to manage SKIRT parameter files with the graphical wizard offered by the MakeUp tool, then you need to install the Qt libraries. In that case, you can just as well use Qt Creator to build the SKIRT code. Alternatively, you can skip graphical user interfaces entirely and perform all tasks on the command line.
• A remote server or multi-node cluster, most likely running a Linux distribution or some proprietary Unix flavor. In this environment you are likely to perform all tasks on the command line, even if it is often possible to use graphical interfaces through a remote windowing system. Installing PTS is only required if you wish to use it on the server for SKIRT-related data preparation or visualization tasks. Alternatively, you can perform these tasks locally and run just the SKIRT simulation remotely.

The installation instructions in this guide are structured in topics as listed below. The procedure for a specific scenario and operating system can hopefully be inferred by picking and choosing from these instructions.

Installation instructions

• Install SKIRT using the command line (Unix or macOS)
• Install SKIRT using Microsoft Visual Studio (Windows)
• Install SKIRT using Qt Creator (all platforms)
• Develop SKIRT using Qt Creator (all platforms)
• Use the SKIRT build options (all platforms)
• Enable multi-processing for SKIRT (Unix or macOS)
• Download SKIRT resource packs (Windows)
• Install a FITS viewer (all platforms)
• Install Python using the command line (Ubuntu)
• Install Python using a graphical interface (macOS)
• Installing PTS (Unix or macOS)

System requirements

SKIRT and PTS can be installed on any recent 64-bit version of Unix, macOS, and Microsoft Windows.

SKIRT

The SKIRT code is intended to be truly cross-platform. It is written in standard C++14, which is supported by all recent C++ compilers, and it uses the CMake build system, which is widely available. The source code for SKIRT is hosted on GitHub. You can download the code as a ZIP archive, but the recommended method is to use a git client, which is widely available. The source code is formatted using clang-format and the documentation on this web site is generated from the source code through Doxygen. You don't need to install these documentation tools unless you are a substantial contributor to the SKIRT project.

The default and most basic build configuration includes a fully-functional version of the SKIRT command-line program with support for multiple parallel threads in a single process. In this configuration, the code has no external dependencies other than the C++ compiler and the CMake build tool. Additional capabilities that may depend on external components can be enabled through user-configurable build options:

• To enable the multi-processing (and hence multi-node) capabilities of SKIRT, the host system must provide an implementation of the standard Message Passing Interface (MPI).
• The optional MakeUp utility, included in the SKIRT code tree, uses the Qt development environment to implement its graphical user interface, so a recent version of the Qt framework must be installed to build this utility.

PTS

PTS is written in Python 3 and requires a Python distribution for language version 3.7 or later to be installed on the host computer. In addition to the functionality offered by the Python standard library packages, PTS also depends on some non-standard but commonly available packages. Usually, these packages can be easily obtained through the Python package manager included with the installed Python distribution. The table below lists the non-standard packages that are used at the time of writing. Note that each of these packages may have additional dependencies, requiring other packages to be installed as well.

Package	Description
python	Python language environment
numpy	General-purpose array-processing and math
scipy	Mathematics and scientific library
matplotlib	Plotting
astropy	Community python library for astronomy
lxml	Support for XML and XSLT
pillow (PIL)	Basic image processing
reportlab	Direct PDF file generator
ipywidgets	Interactive widgets for Jupyter notebook

Once PTS has been properly installed, you can obtain a list of the current (direct) package dependencies by entering the following terminal command:

pts list_dependencies