CPYTOIMPF for qtemp faster than for another library?

I am a programmer working on an IBM AS400 V5R4. I am creating 2 CL programs, both of which use CPYTOIMPF to format the ASF00 PF list for character delimited files. The difference between the two: one formats files in qtemp (PGM A), and the other files in a library other than qtemp (PGM B).

I named program 2 separately to format about 90 files, about 8 million records in total. Both programs consumed more than 1300 s of processor time, while PGM A used about 5% less CPU time than PGM B. I tried several times, and the result is similar.

Can I find out why CPYTOIMPF for qtemp is faster than copying to another library, although there is a slight difference?

+4
source share
5 answers

QTEMP, and this content is not required for survival beyond its service life, so it can be optimized so that it doesnโ€™t transfer things to disk or do it less often, and there may be no need to create file system transactions for the object changes, because if the system goes down The uncontrolled QTEMP is simply reset with the job.

In addition, QTEMP does not need to worry about security controls and objects, since it is a private task.

These things and the like can make access to QTEMP less than a regular permanent library.

+2
source

If you do not get an answer here. Try sending it MIDRANGE-L . There is a large community of IBM experts I (or AS / 400). Probably someone can answer it.

+1
source

QTEMP is always in the iSeries Job library list (with the usual iSeries configured). You have done your - interesting! - experiment with a library that is also on the library list?

I don't know if syslibl, curlibl usrlibl can make a difference.

+1
source

An interesting experiment.

Do you use the SIZE atttribute attribute in the CRTPF command when creating the output file in QTEMP or in another library? I suspect the file extension is faster in QTEMP than in another library. Setting the number of records in the SIZE attribute to the number of records in the input file can reduce the difference in performance because the files no longer need to be expanded.

0
source

In QTEMP , usually you just create a file with CRTFILE(*YES)

The difference may simply be the time a file was deleted before it was recreated, or with a big hit of time to clear the file: MBROPT(*REPLACE)

0
source

All Articles