If you have a many-to-many dimension A with a common dimension B, the storage engine will have to request totals for EVERY member of dimension B from the data measure group when resolving this many-to-many dimension.
It would be VERY nice for performance if, after finding the members of dimension B that exist in your current slice in the intermediate measure group, you could have the storage engine only pull totals for just those members of dimension B from the data measure group.
My specific use case is this... I have a m2m Date and Date Calculations dimension so that a table can drive which days are in YTD 1/19/2011. But these m2m dimensions cause every date partition from the data measure group to be hit even if only recent date partitions are needed for my query.
In an email with Akshai today, he suggested this might be a nice enhancement, if that lends any weight.