diff options
author | jthorn <jthorn@f88db872-0e4f-0410-b76b-b9085cfa78c5> | 2007-01-16 12:47:02 +0000 |
---|---|---|
committer | jthorn <jthorn@f88db872-0e4f-0410-b76b-b9085cfa78c5> | 2007-01-16 12:47:02 +0000 |
commit | 8cb0df53b3d5d615624b9bffdcc11cf13905cfc0 (patch) | |
tree | bcee6a01c349e63d7c957609b5d7b7abeb6b3f36 /test/bl0.75-025/sf_quadrupole_xy[0].xg | |
parent | 2db3b7be3194b3309c74f8674350560fe6cc3e93 (diff) |
set tolerances:
Because AHFinderDirect is solving a nonlinear equation iteratively,
small numerical differences can cause it to make one iteration more
or less, thus changing the solution a little bit. Moreover, the
"equation" is actually a *system* of typically a few thousand equations,
which may be a bit ill-conditioned. The equation tolerance inside
AHFinderDirect is set to 1e-12 for all the test cases, so here we set
the testsuite tolerance to significantly coarser than this, 1e-8.
git-svn-id: http://svn.einsteintoolkit.org/cactus/EinsteinAnalysis/AHFinderDirect/trunk@1500 f88db872-0e4f-0410-b76b-b9085cfa78c5
Diffstat (limited to 'test/bl0.75-025/sf_quadrupole_xy[0].xg')
0 files changed, 0 insertions, 0 deletions