Building ParaView

From ScorecWiki

(Difference between revisions)
Jump to: navigation, search
Revision as of 15:21, 29 March 2006
Duprec (Talk | contribs)

← Previous diff
Revision as of 15:38, 1 June 2007
Zhoum (Talk | contribs)
(Compilation - Serial)
Next diff →
Line 26: Line 26:
$ '''mkdir ''obj-serial'' ''' $ '''mkdir ''obj-serial'' '''
$ '''cd ''obj-serial'' ''' $ '''cd ''obj-serial'' '''
- $ '''cmake ../ParaView'''+ $ '''ccmake ../ParaView'''
 + ''press c to configure''
 + ''press c again to configure''
 + ''press g to generate and quit''
''output truncated'' ''output truncated''
$ '''make''' $ '''make'''

Revision as of 15:38, 1 June 2007

ParaView is a visualization software designed with the need of large data sets in mind. It is able to run on multiple computers using MPI to handle the largest data sets.


What follows is a tutorial to compile ParaView from sources on a Unix computer.

Contents

Prerequisites

  • CMake, version 2.0.6 or newer
  • CVS client to get the sources
  • C++ compiler such as g++
  • MPICH (or other MPI implementation) if parallel ParaView is desired

Getting the sources

We are interested in the very new phasta Reader developped by Kitware, which is (at least for the time being) only available in the development version of ParaView, which means that we need to get the sources from CVS. Replace /space with any directory that has 1GB of available space.

$ cd /space 
$ cvs -d :pserver:anoncvs@www.paraview.org:/cvsroot/ParaView login
Logging in to :pserver:anoncvs@www.paraview.org:2401/cvsroot/ParaView
CVS password: <Press ENTER> 
$ cvs -d :pserver:anoncvs@www.paraview.org:/cvsroot/ParaView co -D "2006-02-13 16:15" ParaView
   output truncated

Compilation - Serial

The recommended way to build ParaView is out of source, meaning that cmake, make etc are not run in the ParaView directory, but in a separate, empty one. This makes it possible to build ParaView for several architectures without collision.

To make the paraview binary, do:

$ mkdir obj-serial 
$ cd obj-serial 
$ ccmake ../ParaView
   press c to configure
   press c again to configure
   press g to generate and quit
   output truncated
$ make
   output truncated

The compilation takes a while. On a 2GHz Pentium 4 running Red Hat Linux 4, it takes about 90 minutes. When it is complete, you have the program ready to use in obj-serial/bin/paraview.

Compilation - Parallel

Compiling parallel ParaView is very similar to the serial case - except that you need to tell cmake where to find the MPI implementation. The following guide assumes a standard SCOREC installation with MPICH in /usr/local/mpich/latest

To make the paraview binary, do:

$ mkdir obj-parallel 
$ cd obj-parallel 
$ cmake -DVTK_USE_MPI:BOOL=ON \
> -DMPI_INCLUDE_PATH:PATH=/usr/local/mpich/latest/ch_p4/include \
> -DMPI_LIBRARY:FILEPATH="-L/usr/local/mpich/latest/ch_p4/lib -lmpich" \
> -DVTK_MPIRUN_EXE:FILEPATH=/usr/local/mpich/latest/ch_p4/bin/mpirun \
> ../ParaView
   output truncated
$ make
   output truncated

When the compilation is complete, the program is ready to use in obj-parallel/bin/paraview.

Personal tools