Changes between Version 39 and Version 40 of InstallUbuntuSvn
- Timestamp:
- Mar 11, 2015, 4:30:25 PM (9 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
InstallUbuntuSvn
v39 v40 1 = Installing current version of ANUGA research code on Ubuntu = 1 = Installing current version of ANUGA research code on Ubuntu 14.04 = 2 3 4 We need {{{subversion}}} to install {{{anuga}}}. 5 6 Use {{{apt-get}}} to install subversion via 7 8 {{{ 9 sudo apt-get install subversion 10 }}} 11 12 Now from your home directory, we can get the {{{anuga}}} source via 13 14 {{{ 15 svn checkout https://anuga.anu.edu.au/svn/anuga/trunk/anuga_core anuga_core 16 }}} 17 18 q 19 20 2 21 3 22 == Packages to install == … … 6 25 7 26 {{{ 8 sudo apt-get install subversion python-dev python-numpy python-scipy python-matplotlib netcdf-bin libnetcdf-dev python-gdal27 sudo apt-get install subversion python-dev python-numpy python-scipy python-matplotlib netcdf-bin libnetcdf-dev libhdf5-serial-dev python-gdal python-pip 9 28 }}} 10 29 … … 19 38 }}} 20 39 21 === Scientific python===40 === Other Python Packages === 22 41 23 We use scientific python to provide netcdf 24 support in python. 25 26 The default python-scientific provided by Ubuntu is 27 version 2.8, and this leads to segmentation 28 faults (at least on 32 bit machines) in Ubuntu 12.04. 29 30 So we need to install scientific python version 2.9.1 from source. 31 Get Scientific python 2.9.1 source. 42 We also need a few other python packages. We can use {{{pip}}} to install these via the command 32 43 33 44 {{{ 34 wget https://sourcesup.renater.fr/frs/download.php/3420/ScientificPython-2.9.1.tar.gz . 45 sudo pip install nose netCDF4 pyproj 35 46 }}} 36 47 48 === Parallel support === 37 49 38 You will need to manually extract via the tar command, i.e. 50 On {{{Linux}}} it is easy to setup parallel support. You can use either {{{mpich2}}} or {{{openmpi}}}. 51 52 Here is the command to install {{{mpich2}}} support 39 53 40 54 {{{ 41 tar xvf ScientificPython-2.9.1.tar.gz 55 sudo apt-get install mpich2 42 56 }}} 43 57 44 You should end up with a directory {{{ScientificPython-2.9.1}}} 58 We interface {{{mpi}}} with {{{Python}}} via the {{{pypar}}} package. 45 59 46 Change into the directory and install scientific python via 60 Download {{{pypar}}} with 47 61 48 62 {{{ 63 svn checkout https://github.com/daleroberts/pypar 64 }}} 65 66 and then change into the {{{pypar}}} directory and install, via 67 {{{ 68 pushd pypar 49 69 python setup.py build 50 70 sudo python setup.py install 71 popd 51 72 }}} 52 73 … … 69 90 This should produce an anuga_core directory in your home directory 70 91 71 === Setup PYTHONPATH ===72 73 We need to tell python where the anuga source code is located.74 This is done via the {{{PYTHONPATH}}} environment variable.75 76 Put the following in your .bashrc file (if you use the bash shell).77 78 {{{79 export PYTHONPATH=/home/******/anuga_core/source80 }}}81 82 where the {{{******}}} should be replaced by the name of your home directory. '''Note the inclusion of {{{/source}}} in the name of the file.''' With some versions of Linux you need to put this command in the file {{{.profile}}}83 84 You should rerun your .bashrc file with85 86 {{{87 source .bashrc88 }}}89 90 Actually it is usually safer to fire up a new terminal to see if you have been successful in setting91 up the PYTHONPATH92 93 You can check your PYTHONPATH with the command94 95 {{{96 printenv PYTHONPATH97 }}}98 99 92 100 93 === Compile ANUGA Code === 101 94 102 Now go to the directory anuga_core and compile the anuga filesvia:95 Now go to the directory anuga_core and install via: 103 96 104 97 {{{ 105 python compile_all.py 98 python setup.py build 99 sudo python setup.py install 106 100 }}} 107 101 … … 111 105 112 106 {{{ 113 python test_all.py107 python runtests.py 114 108 }}} 115 109