Information Technology Reference
In-Depth Information
FIGURE 10.3
Configuration of a WS-PGRADE workflow node.
generator nodes (as shown in FigureĀ 10.2), then it is possible to specify how
to pair the different inputs, using either cross or dot products of the files pro-
duced by the different generator nodes. Finally, if all of the node instances
have finished in the processing phase, the collector nodes receive the results
of the computations and can process them (e.g., they can search for the best
result or can produce some sort of statistics). Note that these phases can be
overlapped or repeated within a workflow; that is, generators and collectors
can be placed into a workflow at any part of the graph without restrictions.
The configuration of workflow nodes can be performed through a simple
interface, as shown in FigureĀ  10.3. It is possible to define the execution
resource, the application, different resource-specific settings, the command
line arguments, and the data used and produced by the job.
Once the configuration of the whole workflow has finished, it can be sub-
mitted. From this point, WS-PGRADE/gUSE is responsible for taking care of
the nodes' execution.
10.3 AutoDock Workflow
This section explains how to use the generic parameter sweep creation
and execution technology described in the previous section for a concrete
Search WWH ::




Custom Search