Given a situation where structured data blocks have been partitioned for analysis on multiple compute nodes, and the data can be saved in EnSight Gold format, such that a set of cases exists which are contiguous from one set to another - the data can be read into EnSight using the casefile Block Continuation capability. This allows any number of these contiguous sets to be clustered and visualized together in an EnSight server. Furthermore, by using this capability combined with EnSight's Server of Servers, one can visualize M number of sets with N number of EnSight Servers. Where m is greater than n.
Each set must be a standalone EnSight case containing a portion of all the parts. The block parts in each of the sets must consist of a valid subset range of each complete block part. This will require, for example, that a given 3D block part will vary in only one dimension throughout the sets. Also, each set must be the neighboring portion or slice in the set of cases.
Example 5: Standalone EnSight Case Containing a Portion of All the Parts
Block Part 1 advances in the i direction.
Block Part 2 advances in the j direction.
Each Case (or set) consists of portions of both parts, which are specified using EnSight's block range capability. Therefore, the first case has the red portions. The second case has the yellow portions, etc.
Each block part can be sliced up in a different direction, but that direction must continue for all sets for that part. And the sets must be the contiguous natural neighbors of each other.
Note: The block dimensions must remain the same in the non-advancing directions, but they can vary per set in the advancing direction (even though, in this simple example, they are quite consistent).
With data set up like the above, we might then decide that we will visualize the
model with only two servers. We could combine (using
Block_Continuation) the first two sets onto the first server and the
last 3 sets onto the second server.
