\section{Compiling the renderer}
To compile Mitsuba, you will need a recent C++ compiler (e.g. GCC 4.1+ or 
Visual Studio 2008+) and some additional libraries, which Mitsuba uses internally. 
Builds on all supported platforms are done using a unified system
based on SCons (\url{http://www.scons.org}), which is a Python-based 
software construction tool. There are some minor differences between operating systems though: for
more details, please refer to one of the next sections depending on which one you use.

\subsection{Common steps}
To get started, you will need to download a recent version of Mitsuba. Make sure that you have the Mercurial (\url{http://mercurial.selenic.com/}) versioning system installed\footnote{On Windows, you might also want the convenient TortoiseHG shell extension (\url{http://tortoisehg.bitbucket.org/}) to run the subsequent steps directly from the Explorer.} and enter the following at the command prompt:
\begin{shell}
$\texttt{\$}$ hg clone https://www.mitsuba-renderer.org/hg/mitsuba
\end{shell}
Afterwards, you will need to download the precompiled dependencies into the
subdirectory \code{mitsuba/dependencies}:
\begin{shell}
$\texttt{\$}$ cd mitsuba
$\texttt{\$}$ hg clone https://www.mitsuba-renderer.org/hg/dependencies
\end{shell}
Common to all platforms is that a build configuration file must be chosen: amongst the
following, please copy the best matching file into a new file to the root of the Mitsuba
directory and rename it into \code{config.py}.
\begin{shell}
build/config-linux.py  
build/config-darwin-x86_64.py  
build/config-darwin-x86.py  
build/config-darwin-universal.py  
build/config-icl11-win32.py  
build/config-icl11-win64.py
build/config-msvc2008-win32.py  
build/config-msvc2008-win64.py
build/config-msvc2010-win32.py  
build/config-msvc2010-win64.py
\end{shell}

Usually, you will not have to make any modification to this fine, but sometimes a few minor 
edits may be necessary. In particular, you might want to add or remove certain
compilation flags from the \code{CXXFLAGS} parameter. The following settings
affect the behavior of Mitsuba:
\begin{description}
\item[\texttt{MTS\_DEBUG}] Enable assertions etc. Usually a good idea, and
enabled by default.
\item[\texttt{MTS\_KD\_DEBUG}] Enable additional checks in the kd-Tree. This
is quite slow and mainly useful to track down bugs when they are suspected.
\item[\texttt{MTS\_KD\_CONSERVE\_MEMORY}] Use less memory for storing geometry (at the cost of speed).
Off by default.
\item[\texttt{SINGLE\_PRECISION}] Do all computation in single precision. This is usually sufficient.
\item[\texttt{DOUBLE\_PRECISION}] Do all computation in double precision. Incompatible with
\texttt{MTS\_SSE}, \texttt{MTS\_HAS\_COHERENT\_RT}, and \texttt{MTS\_DEBUG\_FP}.
\item[\texttt{MTS\_SSE}]Activate optimized SSE routines.
\item[\texttt{MTS\_HAS\_COHERENT\_RT}]Include coherent ray tracing support (depends on \texttt{MTS\_SSE}).
\item[\texttt{MTS\_DEBUG\_FP}]Generated NaNs and overflows will cause floating point exceptions, which can be caught in a debugger. This is slow and mainly meant as a debugging tool for developers.
\item[\texttt{MTS\_GUI\_SOFTWARE\_FALLBACK}]Forces the GUI to use a software
fallback mode, which is considerably slower and removes the realtime preview.
This is useful when running the interface on a remote machine
accessed via the Remote Desktop Protocol (RDP).
\end{description}
All default configurations use the flags \code{MTS\_DEBUG}, \code{SINGLE\_PRECISION}, \code{MTS\_SSE},
as well as \code{MTS\_HAS\_COHERENT\_RT}.

\subsection{Building on Ubuntu Linux}
You'll first need to install a number of dependencies. It is assumed here
that you are using Ubuntu Linux, hence some of the package may be named differently if you are 
using another distribution.

First, run
\begin{shell}
$\text{\$}$ sudo apt-get install build-essential scons qt4-dev-tools scons libpng12-dev libjpeg62-dev libilmbase-dev libopenexr-dev libxerces-c-dev libboost-dev libglewmx1.5-dev libxxf86vm-dev libboost-system-dev libboost-filesystem-dev
\end{shell}
To get COLLADA support, you will also need to install the \texttt{collada-dom} packages or build them from scratch. Here, we install the \code{x86\_64} binaries and development headers available on the Mitsuba website:
\begin{shell}
$\text{\$}$ sudo dpkg --install collada-dom2.2_2.2-2_amd64.deb collada-dom-dev_2.2-2_amd64.deb
\end{shell}
Afterwards, simply run
\begin{shell}
$\text{\$}$ scons
\end{shell}
inside the Mitsuba directory. In the case that you have multiple processors, you might want to parallelize the build by appending \code{-j }\emph{core count} to the command.
If all goes well, SCons should finish successfully within a few minutes:
\begin{shell}
scons: $\texttt{done}$ building targets.
\end{shell}
To be able to run the renderer from the command line, you will also have to import it into your path:
\begin{shell}
$\text{\$}$ . setpath.sh
\end{shell}
(note the period at the beginning -- this assumes that you are using \code{bash}).

\subsection{Building on Fedora Core}
You'll first need to install a number of dependencies. It is assumed here
that you are using Fedora Core, hence some of the package may be named differently if you are 
using another distribution.

First, run
\begin{shell}
$\text{\$}$ yum install mercurial gcc-c++ scons boost-devel qt4-devel OpenEXR-devel xerces-c-devel
\end{shell}
You will also need the \texttt{glew-mx} and \texttt{collada-dom} packages, which are not included in the Fedora package repository. 
You can grab source, \texttt{i386}, and \text{x86\_64} \texttt{RPM} files here: \texttt{http://www.mitsuba-renderer.org/releases}.
Afterwards, simply run
\begin{shell}
$\text{\$}$ scons
\end{shell}
inside the Mitsuba directory. In the case that you have multiple processors, you might want to parallelize the build by appending \code{-j }\emph{core count} to the command.
If all goes well, SCons should finish successfully within a few minutes:
\begin{shell}
scons: $\texttt{done}$ building targets.
\end{shell}
To be able to run the renderer from the command line, you will also have to import it into your path:
\begin{shell}
$\text{\$}$ . setpath.sh
\end{shell}
(note the period at the beginning -- this assumes that you are using \code{bash}).

\subsection{Building on Arch Linux}
There are two ways to install Mitsuba on Archlinux, the Arch way, and the other way.

The Arch Way is to use the Aur software repository. 
Accessing software in the Aur repository is made a lot easier by using a script called packer.

First download packer then use makepkg to build and install it. 
The -is flags will prompt you for your sudo password and then install the package after it has finished building as well as install any needed dependencies.
\begin{shell}
$\text{\$}$ mkdir packer && cd packer
$\text{\$}$ wget http://aur.archlinux.org/packages/packer/packer/PKGBUILD
$\text{\$}$ makepkg -is
\end{shell}

Next use packer to automatically download, build, and install Mitsuba as well as any needed dependencies.
The optional --noedit flag is used if you do not wish to edit the files after they are downloaded.
The optional --noconfirm flag is used if you do not wish to confirm each step of the installation. 
\begin{shell}
$\text{\$}$ sudo packer -S --noedit --noconfirm mitsuba-hg glewmx collada-dom
\end{shell}

Periodically you may wish to update Mitsuba to the latest version.
To do this simply reinstall it and packer will pull and build the latest version.
\begin{shell}
$\text{\$}$ sudo packer -S --noedit --noconfirm mitsuba-hg
\end{shell}

Alternatively you can skip using packer and manually download the files and install them one at a time yourself.
First install glewmx
\begin{shell}
$\text{\$}$ mkdir glewmx && cd glewmx
$\text{\$}$ wget http://aur.archlinux.org/packages/glewmx/glewmx/PKGBUILD
$\text{\$}$ makepkg -is
\end{shell}

And then collada-dom
\begin{shell}
$\text{\$}$ mkdir collada-dom && cd collada-dom
$\text{\$}$ wget http://aur.archlinux.org/packages/collada-dom/collada-dom/PKGBUILD
$\text{\$}$ makepkg -is
\end{shell}

And finally Mitsuba
\begin{shell}
$\text{\$}$ mkdir mitsuba-hg && cd mitsuba-hg
$\text{\$}$ wget http://aur.archlinux.org/packages/mitsuba-hg/mitsuba-hg/PKGBUILD
$\text{\$}$ makepkg -is
\end{shell}

To uninstall do this
\begin{shell}
$\text{\$}$ sudo pacman -R mitsuba-hg glewmx collada-dom
\end{shell}

After installing you will be able to run the renderer from the command line.

If for some reason you are unable access the Aur files, they are also hosted at (\url{https://www.mitsuba-renderer.org/releases/contrib/archlinux/})

\subsection{Building on Windows}
On the Windows platform, Mitsuba already includes most of the dependencies in precompiled form.
You will still need to set up a few things though: first, you need to install Python 
(\url{www.python.org}) and SCons\footnote{Note that on some Windows machines, the SCons installer generates a warning about not finding Python in the registry. In this case, you can instead run \code{python setup.py install} within the source release of SCons.} (\url{http://www.scons.org}) and ensure that they are contained in the \code{\%PATH\%} environment variable so that entering \code{scons} on the command prompt (\code{cmd.exe}) launches the build system.

Next, you will either need to compile Qt 4.$x$ (where $x\ge 7$) from source or grab pre-built binaries from the following page: \url{http://code.google.com/p/qt-msvc-installer}. It is crucial that the build configuration is identical: for instance, when you are compiling Mitsuba using the 64-bit compiler in Visual Studio 2010, you will need Qt binaries made with that same compiler. When compiling Qt from source,
start the correct Visual Studio command prompt, and enter
\begin{shell}
C:\Qt>configure.exe -release -no-webkit -no-phonon -no-phonon-backend -no-script
      -no-scripttools -no-qt3support -no-multimedia -no-ltcg
..(configuration messages)..
C:\Qt>nmake
\end{shell}
inside the Qt directory.

Having installed these dependencies, run the ``Visual Studio \emph{2008/2010} Command 
Prompt'' from the Start Menu (\code{x86} for 32-bit or \code{x64} for 64bit). Afterwards,
navigate to the Mitsuba directory and run \code{scons}. 
Depending on whether or not the Qt binaries are on the \code{\%PATH\%}
environment variable, you may also have to enter
\begin{shell}
set QTDIR=C:\Qt
\end{shell}
where \code{C:$\backslash$Qt} is the path to your Qt installation.
In the case that you have multiple processors, you might want to parallelize the build by appending \code{-j }\emph{core count} to the \code{scons} command.

If all goes well, the build process will finish successfully after a few
minutes. In comparison to the other platforms, you don't have to run the \code{setpath.sh} script at this point. 
All binaries are now located in the \code{dist} directory, and they should be executed directly from there.
\subsubsection{Integration with the Visual Studio interface}
If you plan to modify Mitsuba and you enjoy working within the Visual
Studio interface, the following steps should help you integrate the two
(thanks to Marios Papas for this description!). They are written for Visual Studio 2008,
hence some changes may be necessary on other versions. Also, they assume
that you followed all of the previous steps.

\begin{enumerate}
\item Create a copy of the file \code{config.py} in the Mitsuba root directory and
save it as \code{config\_debug.py}. This configuration 
will later be used to support debug
builds. Open the file and make the following changes:
\begin{enumerate}
\item In \code{CXXFLAGS}, replace \code{/Ox} by \code{/Od} (this disables
optimizations, which is necessary to see variable contents in the debugger)
\item In \code{CXXFLAGS}, add an entry \code{'/Z7'}.
\item In \code{LINKFLAGS}, add an entry \code{'/DEBUG'}.
\end{enumerate}
\item Now, start a fresh instance of Visual Studio and click on 
the \emph{File$\to$New$\to$Project From Existing Code} menu entry.
\item Choose \emph{Visual C++} from the drop-down menu and press \emph{Next}
\item Specify the Mitsuba root directory at the top and enter \emph{mitsuba}
as the project name. 

In \emph{Add files to the project from these folders}, remove the Mitsuba 
root directory and only add the \code{src} and \code{include} subdirectories. 
This is crucial for IntelliSense to work.
Uncheck \emph{Show all files in Solution explorer} and
press \emph{Next}.
\item Choose \emph{Use external build system} and press \emph{Next}.
\item In the following, feel free to replace any occurrence of
\code{scons} by \code{scons -jX}, where \code{X} is the number of
cores in your machine (this will provide faster build times). 
Fill in the following values:

\begin{tabular}{lp{9cm}}
\emph{Build command line}:&\code{scons {-}-cfg=config\_debug.py}\\
\emph{Rebuild command line}:&\code{scons {-}-cfg=config\_debug.py -c \&\& scons {-}-cfg=config\_debug.py}\\
\emph{Clean command line}:&\code{scons {-}-cfg=config\_debug.py -c}\\
\emph{Output}:&\code{dist\textbackslash mtsgui.exe}\\
\emph{Include search paths}:&\code{C:\textbackslash Qt\textbackslash
include} (or wherever the Qt headers are installed on your machine)
\end{tabular}
\item On the next page, uncheck \emph{Same as Debug configuration}.
Afterwards, enter exactly the same values as before except leaving out
the \code{{-}-cfg=config\_debug.py} parameter to \code{scons}.
\item Click \emph{Finish} and wait for IntelliSense to refresh (this might take
a while).
\item Switch to the Release configuration.
\end{enumerate}
\subsection{Building on Mac OS X}
On Mac OS X, you will need to install both scons (\code{www.scons.org}) and 
a recent release of XCode. You will also need to get Qt 4.7.0 or a newer version
--- make sure that you get the normal Cocoa release (i.e. \emph{not} the one based on Carbon). All of the
other dependencies are already included in precompiled form.

Now open a Terminal and run
\begin{shell}
$\text{\$}$ scons
\end{shell}
inside the Mitsuba directory. In the case that you have multiple processors, you might want to parallelize the build by appending \code{-j }\emph{core count} to the command.
If all goes well, SCons should finish successfully within a few minutes:
\begin{shell}
scons: $\texttt{done}$ building targets.
\end{shell}
To be able to run the renderer from the command line, you will have to import it into your path:
\begin{shell}
$\text{\$}$ . setpath.sh
\end{shell}
(note the period at the beginning -- this assumes that you are using \code{bash}).