Are valuable CPU seconds and I/Os being gobbled up by repeated generation of the same indexes ?

Creating temporary indexes repeatedly could be a performance hog. Index generations for larger data base files are rather resource intensive. 

GiAPA offers a report by file or job name that will reveal any larger index generations. If the total time used on index creation for a file is rather high (as in below example where 1 hour 6 minutes were used) then the creation of a permanent index (= logical file) might save quite some resources.