A question about how to code a procedure to use dbms_parallel_execute package for a huge insert

Hi Experts,

I am on Oracle 11.2.0.4 on Linux. I have a query to tune and  a scenario in which the query executes fine for smaller data sets and takes more time for huge data set - when it inserts more than one million rows. Our application timeout is 60 seconds and when the data inserts is more than 1 million we are getting timeout.  I want to try out dbms_parallel_execute package for this - to change that query to use this package. This question is specifically about this requirement. I am asking this question to specifically know about how to use DBMS_PARALLEL_EXECUTE package to make an insert go in parallel. (Please do not spend your time on other aspects of tuning as I will raise a new question for it if I need help on that.)