When loading a VTK Multiblock (*.vtm) I have the option of a dropdown box that lets me to choose Block or Interleave. What exactly does this do?
If each block is composed of multiple pieces, does it distribute the pieces when I choose Interleave? that is, distribute by piece then by block?
If I choose Block, does it distribute by Block first, then by pieces if they exist?
“Piece Distribution” does not seem to occur in ParaViewGuide-5.8.0.
Mouseover help just says “Control how datasets are loaded in parallel.”
Let *X* be the rank of a specific reader, and *N* be the number of reader, then:
* `Block` Each processor is assigned a contiguous block of files, [ *X* * *N* , ( *X* + 1) * *N* ).
* `Interleave` The files are interleaved across readers, *i* * *N* + *X* .
It doesn’t make any distinction between blocks and pieces. All leaf nodes in the multiblock dataset hierarchy are treated the same.