For only 3M rows, it depends on what the values of x and y are. If the difference between x and y is small, then you would probably be better off in with an index. The index will perform a range scan between x and y.

If the difference between x and y is great (1 month) then a full partition scan would take approximately the same time as a index range scan + table accesses. You would get similar performance either way.

That being said, there are other advantages to partitioning, especially in your case. If you are going to periodically purge this data based on the month it was created, the ease of management would definitely be a reason to choose partitioning.