ASO Cube with attributes very slow in retrieval

Question

Hi,
I have a ASO Cube with 5 base dimensions and 8-9 attributes on the entity dimension. I have only 5-6 measures, which do the averages and counts based on the 40 day period. Howere, the data is loaded at the 15 minute increment
Entity
Date - (date-time, lowest level being date)
TIme - ( 15 minute time for the full 24 hour period, has a attribute assocuated with oit)
LocationType
Measures.
The sample formula is
IIF(Islevel([Locations].CurrentMember,0), Avg(CrossJoin({[Measure].[Sale]},{[DateDim].CurrentMember.lag(40):[DateDim].CurrentMember})),Missing)
Is there a way, i can have this calculated as a part of the script? DO you suggest i create a BSO, to do these calculations and pass on the result.
In OBIEE, the report is to display the followung based on the date input.
Entity Gen7, Entity Gen 6..... Entity Gen 2, Attr1, Attr2, Attr3, Attr4, Attr5, Attr6, Attr7, Measures

Answer 1

2 things I would look at
1st - I don't know how much performance you would get out of this, but I'm not clear why you are using a crossjoin in your MDX, it seems unnecessary and may cause more overhead. The following should work, you could also try using IsLeaf instead and see if that is any faster
IIF(Isleaf([Locations].CurrentMember),
Avg({[DateDim].CurrentMember.lag(40):[DateDim].CurrentMember},
[Measure].[Sale], INCLUDEEMPTY)
2nd - your problem mostly revolves around the fact that you are running a 40 member sum/avg for every member you are querying. It also sounds like the average is at the Day level, which is not level 0. So for all forty days, ASO also has to calc the results of each of those days. Remember that aggregations only get you so far, you should really think of everything in ASO as dynamic and that is why you can see what you have set up is not going to work that well, it is too calc intensive.
I don't know how practical this is, but to get this to work fast you would probably need to break out the 15 minute increments below the day level to another dimension so the day level becomes a stored level zero member. The 15 minute increment dimension should also be stored. If at all possible you would want to have an alternate stored hierarchy with the 40 days you want to base the average on. Enable alternate hierarchies in your aggregations, then change your MDX calc to be based on the parent of the 40 day hierarchy divided by 40. That would be fast.
I suppose you could opt to not break out the 15 minute increments and just have the shared hierarchy made up of the 15 minute increments that are below the 40 days. That would still give you a good stored subtotal that with some query hints you could get optimized.