[visit-developers] LLNL-VisIt Testing: Message(s) from VisIt's Automatic Test Run

Brad Whitlock whitlock2 at llnl.gov
Thu May 7 14:04:26 EDT 2009

Does that require domain adjacency? I don't think CGNS serves that up either.


>Brad Whitlock wrote:
>>CGNS.py should not necessarily be rebaselined. I think the 
>>differences relate to streamlines across domain boundaries but they 
>>are not crossing the boundaries. It makes sense; I don't think CGNS 
>>serves up ghost data currently. If this is what's happening, we can 
>>either rebaseline or I can try making CGNS serve up ghost data.
>Dave will have to respond here, but I believe the new parallel 
>streamlines algorithm does not require ghost zones.  If there are no 
>ghost zones, streamlines are integrated until they barely exit the 
>current domain, and are then sent to the neighboring domain for 
>continuation.  So they should no longer stop at domain boundaries, 
>even if there are no ghosts.
>Sean Ahern
>Oak Ridge National Laboratory
>AIM: ornlsean

Brad Whitlock                   Lawrence Livermore National Laboratory
whitlock2 at llnl.gov

More information about the visit-developers mailing list