You may or may not need partitioning.

But you absolutely need to determine the proper clustered index for the table.

It's the column(s) you (almost) always specify in a WHERE clause, particularly if they are in a range.

In your case, it's likely "startdttm", although I can't say for sure without more details.

That will give you vastly better performance, partitioned or not.

Then, if you do partition, partition on the clustering key. Partitioning does allow you to specify different compression for each partition, if you want to (i.e., you can compress older, historical data w/o being forced to compress current data, even though both are in the same table). Sadly, I don't think we can yet specify a different FILLFACTOR for each partition, which could be tremendously helpful as well.