Compared to the single-thread approach of SQL Server itself, SQL Parallel Boost facilitates the parallel execution of any data modification operations (UPDATE, INSERT, DELETE) - making best use of all available CPU resources. This results in performance gains of up to factor 10 (!), which is typically useful in data warehouse and ETL environments, or situations you're facing massive data modification operations. SQL Parallel Boost is purely based on T-SQL. The built-in dispatching function automatically splits any statement into activities, which are the executed in parallel. The currently supported splitting rules are: even, or no split if count of rows affected is below a predefined threshold. The parallel execution can be tuned thru job related priority settings. Potential deadlocking situations are handled automatically (retry) and the execution of queued (waiting) activities can be indirectly cancelled thru optional timeout values.