next up previous index
Next: 8 Task Level TCAD Up: 7 Data Management Previous: 7.1.5 Persistence and Robustness

7.2 Splitting and Data Dependency

 

During multi-step simulation, a step uses its predecessor's output to generate output on its own. Therefore, all data produced by a step depend on the data produced by previous step. For split-experiments, all runs - and all data of all pertaining steps - that branch after a given step depend on the the parent's step data. Figure 7.1 shows three runs created from a flow with three steps, with runs number 2 and 3 splitting from run 1 after step 1. If run 1 is removed, runs 2 and 3 are said to be orphaned,   indicating the fact that their split parent is not available any more .   Note that only data from the second step on are available in runs 2 and 3.

 figure1938
Figure 7.1:   Runs 2 and 3 split from run 1 after step 1. Data for step 1 is only available in run 1. After removing run 1, runs 2 and 3 are orphaned. Information about how step 1 was produced is maintained, but output data are not available.



Christoph Pichler
Thu Mar 13 14:30:47 MET 1997