ibmi-brunch-learn

Announcement

Collapse
No announcement yet.

CPU % only speed up at first run.

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • CPU % only speed up at first run.

    Hi, I am in performance test. I tried SQL indexing in my last test. When I tried it, I submit the jobs, every job CPU % has 7-9% (my program run in n-way). After 20 seconds I checked the record and 1 job already proceed 1800 record (the previous test only around 800 records in 20 seconds). Then I leave the process and do my other task.
    After 20 minutes I checked the n-way process and the records proceed become slower (almost same as the previous test). Then I checked the job. Every jobs only has CPU % in 0.1%.

    is it normal or there is some setting I missed ?

  • #2
    There is practically no point in checking CPU% during the first 10 to 20 seconds of a long running job. It will regularly indicate that much more CPU has been used than if you check a minute or two later, simply because much more CPU was used. The startup of most jobs can take more CPU than perhaps the rest of the jobs' life. A whole bunch of work goes into preparing a job's environment. The closer to the job start, the higher the number you're likely to see, approaching 100% in the right circumstances.

    If you want to review a job's run-time performance with WRKACTJOB, try F10='Restart statistics' after the first 10-20 seconds have gone by. Then check again with F5='Refresh' a few minutes later and at intervals after that. At least a couple minutes between checks is needed in order to gather enough info.

    By restarting statistics, the startup activity will be eliminated from the cumulative values, and you'll mostly be seeing only actual run-time numbers (which will usually be much lower).

    That is separate and very different from any measures that might (or might not) indicate how many rows of a table have been "processed". Some details of any table can be needed before numbers can be understood. Factors such as locations of deleted rows or physical sequence, among others, may cause radical variations.
    Tom

    There are only two hard things in Computer Science: cache invalidation, naming things and off-by-one errors.

    Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?

    Comment


    • #3
      Thanks tom for the explanation

      Comment

      Working...
      X