sunrise-for-tipsy

Sunrise is a popular "Monte-Carlo Radiation Transfer code for calculating absorption and scattering of light in astrophysical situations". Unfortunately, it is notoriously finnicky to get installed, as it relies on specific point release versions of nearly a dozen different libraries. Ben Keller introduces a way to build Sunrise on Docker in his blog. However, his instruction require root permission which does not work for people who is working on the server. By referring Sunrise's wiki and Ben Keller's blog, this instruction will build Sunrise as a normal user at your home directory on server.

Prerequisites

ICC compiler

Unfortunately, sunrise will not compile with any free compilers. You will need to get the Intel C++ compil on your server. If the ICC compiler does not work on your server, you can get it from intel and you may need to pay for it though and you have to consume your precious time on installing it. If you do not want to spend time on this step, please contact the administrator to solve this problem.

Set up Mercurial

Mercurial is a free, distribute source control management tool like git. In the following instructions, you need it to clone projects rather than git so that Mercurial is highly recommended to install.

You can find full descriptions for Mercurial setting up on Mercurial's webpage. Since Mercurial is written in Python, it is much easier to get Mercurial working on your server by using pip or conda as a result, just execute

pip install Mercurial

or

conda install -c conda-forge mercurial

This shoudl work on all platform of Linux, even on OSX and Windows. If you meet the permission problem by using pip, you can add the option --user to have a try.

Getting the libraries

Now we need to begin by fetching all of the varied specific versions of libraries that sunrise relies on.

Boost

Download Boost from SourceForge, note that Sunrise requires the version 1.48.0 or later, and version 1.48.0 is confirmed to work, later version may cause problems. You can fetch version 1.48.0 from boost archive. Cd to the directory where you put the distribution, then patch this ancient version of boost so it compiles with a modern compiler:

wget --no-check-certificate https://svn.boost.org/trac/boost/raw-attachment/ticket/6165/libstdcpp3.hpp.patch
patch -p0 < libstdcpp3.hpp.patch
for file in `grep -r -l TIME_UTC *`; do sed -i -e 's/TIME_UTC/TIME_UTC_/' $file; done

Build it:

sed -i -e 's/using intel-linux/using intel-linux : : : <compileflags>-shared-intel <linkflags>-shared-intel/' project-config.jam
echo "using mpi : /opt/intel/impi/5.0.3.048/intel64/bin/mpicc ;" > user-config.jam
./bootstrap.sh --without-libraries=python --with-toolset=intel-linux --prefix=$HOME --libdir=$HOME/lib/boost_1_48_0 --includedir=$HOME/include/boost_1_48_0
./b2 -j<n> --user-config=user-config.jam install

This will copy the Boost headers to $HOME/include/boost_1_48_0/boost and will build the libraries and put them in $HOME/lib/boost_1_48_0, you can change them to any directories you want to put the headers and libraries at the options of ./bootstrap.sh. Also, <n> should be some reasonable number depending on how many cores your machine has, because building Boost in parallel greatly decreases the build time.

If you use the above commands, Boost will install its include files in a directory $PREFIX/include/boost-<version>/boost. For the files to be found without specifically adding that directory to the include directives, you can do

ln -s $PREFIX/include/boost-<version>/boost $PREFIX/include/boost.

Blitz

First clone the Mercurial repository and checkout the ab84372f3dce branch which is confirmed to work:

hg clone http://blitz.hg.sourceforge.net/hgweb/blitz/blitz
cd blitz/
hg checkout ab84372f3dce

Build it:

autoreconf -fiv
export CXX=icpc
export CC=icc
export CXXFLAGS='-O2 -g -pthread -shared-intel'
export CFLAGS='-O2 -pthread -shared-intel'
export LDFLAGS='-pthread -shared-intel'
 ./configure --prefix=$HOME --enable-threadsafe --disable-cxx-flags-preset --enable-serialization 
make lib
make install

If you get errors related to not finding the boost libraries try using --with-boost-libdir=$HOME/lib/boost_1_48_0/ (or wherever you installed boost).

cfitsio

CFITSIO is a library of C and Fortran subroutines for reading and writing data files in FITS (Flexible Image Transport System) data format. You can find the latest version at NASA software archive and fetch it:

wget http://heasarc.gsfc.nasa.gov/FTP/software/fitsio/c/cfitsio_latest.tar.gz

Untar the file and do:

export CFLAGS='-pthread -O3'
export CC=icc
./configure --prefix=$HOME
make
make install

CCfits

CCfits is an object oriented interface to the cfitsio library. It is designed to make the capabilities of cfitsio available to programmers working in C++. It is written in ANSI C++ and implemented using the C++ Standard Library with namespaces, exception handling, and member template functions. The download link is on NASA sofrware archive and fetch it:

wget https://heasarc.gsfc.nasa.gov/fitsio/CCfits/CCfits-2.5.tar.gz

Extract sources to a directory named CCfits (otherwise there will be trouble..)

export CXXFLAGS='-pthread -O3'
./configure --prefix=$HOME --with-cfitsio=$HOME
make
make install

libPJutil

First get libPJutil from bitbucket by doing

hg clone https://bitbucket.org/lutorm/libpjutil

It is usually fine to use the most recent version of this, as it doesn't change much. Then go into this directory and do:

autoreconf -fiv
export CPPFLAGS=-I$HOME/include (or wherever you need it to point to find blitz; also use "-pthread" if you use it for blitz)
export CXXFLAGS="-O2 -pthread"
export LDFLAGS=-L$HOME/lib (or wherever)
./configure --prefix=$HOME  (if boost is in $HOME/include and $HOME/lib)
make
make install

This library now contains an embedded copy of the GNU "units" program for units conversion. For this to work, the path to the file "units.dat" needs to be defined in the environment variable UNITSFILE. On standard linux systems, this file is in /usr/share/, so in that case you would add

export UNITSFILE=/usr/share/units.dat

to your bash_profile. If you don't have GNU units installed, you need to download it and put this file somewhere. NB. As of v2.0, this file is no longer included, so you must download v1.88.

Building Sunrise

Get sources from the repository with git:

git clone https://bitbucket.org/lutorm/sunrise.git

In general you should check out the current version, which is mentioned on the front page, not the trunk. The trunk may be unstable at any given time. Then build it:

autoreconf -fiv
export CPPFLAGS="-I$HOME/include -I$HOME/include/libPJutil" (yes, you need to separately specify the libPJutil directory, sorry)
export CFLAGS="-O2 -pthread"
export CXXFLAGS="-ggdb  -O3  -pthread"
export LDFLAGS="-pthread -L$HOME/lib"
export CXX=icpc
./configure --prefix=$HOME --with-boost=$HOME --with-boost-thread=mt --with-boost-program-options=mt

Finally, it's time to compile:

make
make install

If you find the executables sfrhist, mcrx and broadband in the subdirectory src, congralutions, you have done to make Sunrise work on your server to produce mock observational images from Tipsy format simulation data.

Making mock observations

To make a mock observation images from the snapshots with tipsy units (e.g. Gasoline/ChaNGa), You should install pynbody and compile smooth and std2ascii on your server.

pynbody

Pynbody is a light-weight, portable, format-transparent analysis framework for N-body and hydrodynamic astrophysical simulations supporting PKDGRAV/Gasoline, Gadget, N-Chilada, and RAMSES AMR outputs. You can find its installation guide and tutorials on its website.

smooth

You can find smooth on N-bodyShop's GitHub page, and fetch it by doing:

git clone https://github.com/N-BodyShop/smooth

to build:

make

The smooth executable should exist in the directory.

std2ascii

std2ascii is one executable in tipsy_tools on N-bodyShop's GitHub page, and fetch it by doing:

git clone https://github.com/N-BodyShop/tipsy_tools

Before build the whole set, remember to add std2ascii to TOOLS variable

TOOLS = ascii2bin bin2ascii totipnat totipstd snapshot std2ascii

and add a line for std2ascii

td2ascii: std2ascii.o
        $(CC) $(CFLAGS) -o std2ascii std2ascii.o -lm $(LIBS)

as std2ascii is not the default executable produced by tipsy_tools. Then build it:

make

The std2ascii associated with other five executables are in the directory.

Scripts producing mock observations

It is the time to fetch scripts which created by greg stinson who hacked away at converting old idl scripts to python to produce your mock observations. I recommend you clone this repository to your local directory and make symbolic links for all scripts in the directory you want to make an image of and move the executables sfthist, mcrx, broadband, smooth and std2ascii to an identical directory, then change the path to the variable SR in sunrise.sh.

If you have everything well installed, make sunrise images is as easy as these one line:

nohup ./sunrise.sh &

or if you want to align the coordinates with the spin of stars, just doing

nohup ./sunrise.sh -s &

The images will be called edge.png and face.png. Enjog!

Troubleshooting

If the installation fails with the configuration with the message

configure: error: Could not find a version of the boost_system library!

you can also use --with-boost-libdir=$Home/lib/boost_1_48_0 (or whatever) to help the compiler find the libraries.

After successfully compiling the boost,if you get the error

error while loading shared libraries: libboost_serialization.so.1.48.0: cannot open shared object file: No such file or directory

you have to add the path to the boost libraries to the environment variable LD_LIBRARY_PATH by export the path like this:

export LD_LIBRARY_PATH=$Prefix/lib/boost_1_48_0:$LD_LIBRARY_PATH

Replace $Prefix/lib/boost_1_48_0 with the path you choose to install boost libraries to.