Commits

Thu Sep 13 21:07:32 UTC 2001

(1) Added support for writing POOMA R1 compatible DiscField fileSets. 
This implementation is simplified in that it writes only a single 
fileSet from context 0. However, it works with both Arrays and 
multi-material/multi-centering Fields. The latter objects must fit 
entirely inside a record. This implementation is designed to work in 
parallel, but we haven't tested it on Nirvana. Once we verify correct 
parallel operation, I will put this on the Blanca branch. Jim is working 
on the reader and that is nearly done as well.

(2) Added a class called CollectFromContexts that accumulates values 
from all contexts into an std::vector on a specified context. This data 
can be processed and then broadcast back to the other contexts.

(3) Added a new compressed() free function, which is used to see if an 
engine is fully compressed. Added support for this to both Fields and 
Arrays.

(4) Made it possible to assign a Field with 1 sub-field to an Array.

(5) Added free functions---numMaterials(x), centeringSize(x), 
subField(x, m, c)---where x is either a Field or Array. These adapters 
allow one to write code that will work with both Arrays and Fields.

This was tested using CodeWarrior Pro 7 under MacOS X. It was reviewed 
by Jim Crotinger.

I will do further testing under Linux with GCC and KCC and fix any 
problems I encounter.

Scott

<Attachment missing>
-------------- next part --------------
(1) Added support for writing POOMA R1 compatible DiscField fileSets. 
This implementation is simplified in that it writes only a single 
fileSet from context 0. However, it works with both Arrays and 
multi-material/multi-centering Fields. The latter objects must fit 
entirely inside a record. This implementation is designed to work in 
parallel, but we haven't tested it on Nirvana. Once we verify correct 
parallel operation, I will put this on the Blanca branch. Jim is working 
on the reader and that is nearly done as well.