Performance Tuning :: Getting Alternative To REPLACE Function?

I have a table with:

1 million rows
average row length 200 bytes
50 columns
and this update statement
UPDATE mytable SET varchar2_4000_column = replace(replace(replace...300 times)

It looks at every row in the table (no WHERE clause) and does these 300 replace operations on this column for each row. Each replace replaces with a null so effectively it is removing strings. Much of the time these strings are not in the column.

This update statement takes 25 minutes and it is 98% CPU and 2% USER_IO time.

I figure that is what is taking all the time since it is a CPU bound statement. if rows in this table are persistent over time then tag rows with a flag to show which ones have already been processed and skip these next time around.