Describe Data Parallel Model?

In data parallel model most of parallel work concentrates on performing operations on a data set. Data set is characteristically organised in a common structure like an cube or a array. A set of tasks work together on same data structure with every task working on a different parts of same data structure. Tasks perform same operation on their part of work E.g. "add 3 to each array element" can be one task. In shared memory architectures all tasks can have access to data structure via global memory. In distributed memory architectures data structure is split up and data inhabits as 'chunks' in local memory of every task.