[Octopus-notify] svn commit: r2816 - trunk/testsuite/finite_systems_3d by acastro

[octopus-notify at tddft.org]

      Author: acastro
        Date: Thu Apr  5 23:49:06 2007
New Revision: 2816
   Changeset: http://www.tddft.org/trac/octopus/changeset/2816

Modified:
   trunk/testsuite/finite_systems_3d/18-current.01-current.inp

Log:
Finally I figured out why this test was failing. It seems that, due
to rounding differences, the PoissonSolver (cg in this case) was stopping
one iteration before or after depending on which compiler or compiler flags
one is using. Since the final number that is checked in the test is actually
rather small (magnetic field created at the center of a loop of electronic
current) this led to a small difference that was captured by the test.

By decreasing PoissonSolverThreshold a little bit one gets almost exactly
the same results.