Is table partitioning required for parallel query processing?
No, queries against unpartitioned tables can benefit from parallelism, especially for deep joins. Partitioned tables, where the partitions are on separate devices, however, are likely to provide the best performance for parallel queries. Data can then be retrieved from the different partitions simultaneously, assuming there is one worker process for each partition. Creating a clustered index on a partitioned table also provides benefits-multiple insertion points and faster retrievals.
