(Discussion) Performance impact on the processing by defining aggregates in FM

[Desperado]

What will be the different when i create all my calculations in the FM rather than creating in Report Studio ? I heard that there will be a performance impact if you do it in FM .... my doubt is how will it improve the performance ?
Thanks in Advance

[goose]

Thats a very interesting question.... I would test the scenario first wrt a performance gain. The benefit of having calcs in FM is that they are available globally for report development BUT there are also limitations as to what calcs you can do, for example aggregation of calcs that have division dont always work...

[blom0344]

In the most general sense, defining operations in the backend will give the most gain.
The average back-end will consist of a full-size server with loads of memory and is very well suited for performing calculations
However, your post title refers to ' defining aggregates'  in FM. The most common scenario is that facts are 'rolled-up'  by means of the aggregation behavior. In most cases this will translate to an aggregate SQL statement which is handled by the database server you report against in most cases anyway.

More complex query request may involve additional input from the Cognos server, which is some cases is beneficial and in others not.
As Angus suggested, create a test case and measure performance..