With a bit more information, we maybe can be more help...
What is the source of the data?
Approximately how many rows are you processing during a run?
What settings are you using on the Update step?
What is the target database (PostgreSQL, SQL Server, mySQL, whatever)?
Do all rows you are processing require updating (or are some entries unchanged)?
If some/many entries are unchanged, you could try the Merge/Synchronize after merge approach (that is sometimes vastly faster for us, but it depends heavily on the nature of the data that's being updated).