Performance of CA Allocate

Document ID : KB000049846
Last Modified Date : 14/02/2018
Show Technical Document Details

Description:

Vkgparms to specify to provide performance improvement. How to determine CPU usage of Allocate which varies with the amount of ASRs.

Solution:

CA Allocate does have some overhead. The standard "line" I used when I did training was:

For any beneficial "work" to be done, there is some price to pay.

The amount of overhead of CA Allocate is different for every user. The overhead is a factor of:

  • How much ASRs you are running.
  • The number of volumes defined in the VDSTORGP member

20,000 lines of ASRs will have much more overhead (CPU time used) than 2,000.

A test you can perform to get an idea of the amount of overhead your ASRs cause is to setup a string of IEFBR14's - 100 or more in the string.

First, run the string with VAM active and note the job and CPU time.

Then, run the string so VAM is bypassed. Either add a statement in the ASRs to "EXIT CODE(0)" for your jobname, or add the //VDSBYPAS DD DUMMY in the step. Then note the job and CPU time and see the difference.

Changes have been made to improve the performance of CA Allocate. In early 2008, maintenance was delivered that improved Allocate performance significantly. (This logic is in r12.5 "base".)

Tests by a major banking firm in New York, in a controlled "lab" environment, showed that before this maintenance, CA Allocate performance was notably inferior to a competitor product.

With the maintenance applied and the new Vkgparms activated, the results showed that Allocate performance was better than "them" by:

  • 13% faster in elapsed time
  • 17% less MIPS
  • 19% less CPU time
  • Reduction in Allocate elapsed time was 50% !

(The tests were done using IDCAMS allocations of 11,591 data sets.)

This site performs upwards of 150,000 allocations per hour, so, performance is paramount. With Allocate, you have the best performing allocation control product available!

In a recent call to my contact at this site, he states he has had no problems with this maintenance and has no uncertainty about recommending activating this logic.

To get this performance boost, specify these Vkgparms:

PLSOPT79 (Y)
PLSOPT80 (Y)
PLSOPT81 (Y)
PLSOPT82 (Y)

Then either restart VAM or do this command:

F vam,REFRESH,PARMREF

If you have any questions, please contact CA Technical support.