[Octopus-devel] cvs commit: marques/octopus/testsuite/periodic_systems_1d 01-free_electrons.01-ground_state.inp 01-free_electrons.test 02-cosine_potential.01-ground_state.inp 02-cosine_potential.test 03-sodium_chain.01-ground_state.inp 03-sodium_chain.test Makefile.am cosine_potential.test free_electrons.test sodium_chain.test

[cvsusers at tddft.org]

appel       2006/03/11 20:26:30 WET

  Added files:
    octopus/testsuite/periodic_systems_1d 
                                          01-free_electrons.01-ground_state.inp 
                                          01-free_electrons.test 
                                          02-cosine_potential.01-ground_state.inp 
                                          02-cosine_potential.test 
                                          03-sodium_chain.01-ground_state.inp 
                                          03-sodium_chain.test 
                                          Makefile.am 
  Removed files:
    octopus/testsuite/periodic_systems_1d cosine_potential.test 
                                          free_electrons.test 
                                          sodium_chain.test 
  Log:
   The testsuite has now a new file naming scheme. Input files are
   stored in separate files and are not contained in the .test files
   anymore. Also the RUN keyword disappeared. Instead a line like
  
    Input: 01-carbon_atom.01-ground_state.inp
  
   specifies the input file and implies a run at the same time. I also
   added DebugLevel = 0 to all input files for ease of typing while
   debugging a test.
  
   It is now also possible to commit tests for parallel runs. This
   is indicated by the executable for which a test is valid (e.g.
   octopus_cmplx_mpi). With the keyword 'Processors' one can determine
   how many processors a given 'Input' line requires. The 'Processors'
   keyword has to be given always before 'Input'. If none is specified
   the value of the last appearance or if not available a default of
   2 processors is taken. As example:
  
   # specify on how many processors we should run this test
   Processors : 3
   Input      : 02-neon_mpi.01-ground_state-kli_par_domains.inp
  
   Match ; TotalEnergy ; grep -A 6 '^Energy:' static/info ; \s*Total\s*=\s*-28.1938
  
   # now we start a td calculation parallel in states on 5 processors
   Processors : 5
   Input      : 02-neon_mpi.02-td_calc-kli_par_states.inp
  
   # another run, also with 5 processors (note we don't override the
   # previous value of 'Processors')
   Input      : 02-neon_mpi.03-td_calc .....
  
   ...
  
   This allows e.g. to test restarts with a different number of processors.