Does pipeline parallism exist in server jobs?

zulfi123786 · Post by **zulfi123786** » Wed Jun 10, 2009 12:16 pm

chulett · Post by **chulett** » Wed Jun 10, 2009 12:55 pm

bobyon · Post by **bobyon** » Wed Jun 10, 2009 1:05 pm

Dang, Finally a question I could have answered and you beat me to it.

chulett · Post by **chulett** » Wed Jun 10, 2009 1:09 pm

You could always go ahead and answer the inevitable next question.

IBM Analytics Champion 2009 - 2020 · Post by **asorrell** » Wed Jun 10, 2009 1:48 pm

Actually - I am going to ask the inevitable question... "how?"

I was aware that parititioned parallelism could be implemented manually using link partitioners and link collectors. However I thought that server processed the data through the stages one step at a time - preventing pipeline parallelism.

ddevdutt · Post by **ddevdutt** » Wed Jun 10, 2009 2:02 pm

You can achieve pipeline parallelism by using the IPC stage or by turning on inter process row buffering.

asorrell wrote:Actually - I am going to ask the inevitable question... "how?"

I was aware that parititioned parallelism could be implemented manually using link partitioners and link collectors. However I thought that server processed the data through the stages one step at a time - preventing pipeline parallelism.

chulett · Post by **chulett** » Wed Jun 10, 2009 2:18 pm

ddevdutt wrote:You can achieve pipeline parallelism by using the IPC stage or by turning on inter process row buffering.

We have a winner! :D