home page

Installing Cactus

  • Get the GetCactus script from the Cactus website.
  • Make the flesh checkout and then the arrangements checkout.
  • Get AEIDevelopment arrangement via CVS. The command is
    cvs -d :pserver:<user>@cvs.aei.mpg.de:/numrelcvs login
    cvs -d :pserver:<user>@cvs.aei.mpg.de:/numrelcvs co AEIDevelopment
  • Get AEIThorns arrangement
    cvs -d :pserver:cvs_anon@cvs.aei.mpg.de:/numrelcvs login
    cvs -d :pserver:cvs_anon@cvs.aei.mpg.de:/numrelcvs co AEIThorns
  • Get LSUThorns and LSUDevelopment
    cvs -d :pserver:<user>@cvs.cct.lsu.edu:/numrel login
    cvs -d :pserver:<user>@cvs.cct.lsu.edu:/numrel co LSUThorns
    cvs -d :pserver:<user>@cvs.cct.lsu.edu:/numrel co LSUDevelopment
  • Get LSU_HYPER
    cvs -d :ext:<user>@relativity.phys.lsu.edu:/home/perturbed/cvsroot co LSU_HYPER
  • Get also TAT, AlphaThorns
    cvs -d :pserver:cvs_anon@cvs.cactuscode.org:/arrangements co TAT
    cvs -d :pserver:cvs_anon@cvs.cactuscode.org:/cactusdevcvs co AlphaThorns
  • Get Carpet
    Visit the Carpet website for more details. The following command was taken from there. It gets you the development version of Carpet via wget.
    wget -r -nH -np --cut-dirs=1 -R "index.html*" http://www.carpetcode.org/\~darcs/carpet/
    After this, make softlinks for the different Carpet thorns in the arrangements directory.

    Compilling the code

  • Set up your .bash_profile. Here is the one I use.
  • You need a thorn list and a parameter file. I use to have some here but they change from time to time. Go to the Cactus WaveToy Demo to get those two files.
  • A configuration options file. This depends on the machine architecture and the paths of the diferent source code and libraries. This one works well with my IBM T42. For a more generic file look here.

  • Very well, time to compile. Type the following
    gmake hyperGR-config THORNLIST=ThonList.th options=config
    gmake hyperGR
    Here, hyperGR is just a name for the executable, ThornList.th and config are the thornlist and configuration files you got from the Cactus website.

    Running!

    Finally, you need a parameter file. Again, you can get one from the Cactus website. To run the code in one procesor type
    mpirun -np 1 ./cactus_hyperGR parfile.par

    Functional copy

    If you need Cactus in another machine (a cluster, for instance) but you want just the code, this command can be useful

    rsync -avz <Cactus dir>/ --exclude test/ --exclude doc/ --exclude configs/ --exclude exe/ --exclude par/ --exclude _darcs/ --cvs-exclude user@machine:/<dir>/

    It helps to keep all your copies of the Cactus tree in different machines updated.

    BEWARE!!! If you do this, testsuites don't run properly.




    Enrique Pazos
    Last modified: Fri May 11 16:35:28 CDT 2007