QUERY_MAX_WP_DIAG parameter for parallelization changed but no performance optimizitaion

[regysm21]

Hi experts,

we did some tests and increased and reduced the number of parallel processes for processing Multiprovider sub-queries through the parameter QUERY_MAX_WP_DIAG in the table RSADMIN from 5 to 2 and then 10 and there is no performance improvance in the query. The query has the same runtime. When we increase to 10 the WP, it uses 5 WP and when we reduce to 2, it uses 2 then. But still no performance difference...

Can anyone explain this please?

Best Regards,

Regys

[FCI]

It could mean that one of the granular task is much longer than the others. In this case, whatever the number of processes is, the total duration is always equal to the duration of this task.

Regards,

Frederic

[regysm21]

Hi Frederic,

thank you for the answer.. I also thought the same for it but i was wondering if there is some other factor that may prevent it...

Best regards,

Regys

[former_member186445]

and your system specifications are?

[regysm21]

Hi Mario,

thanks for the reply. I tried this in a non-HANA BW and HANA BW with 19 workprocesses each. (10 Dialog Processes, 6 Batch, Spool 1, Update (UPD) 2).

I understand that if I increase the Parameter at Query_MAX_WP_DIAG it cannot use more than 6 parallell processes, but what I don't understand is that why the query runtime remains the same if I use only 1, 2 or 6 Workprocesses?

Best regards,

Regys