Question on cube build / query performance (11.2.0.3)
Hi, I have a dumb question on cube build performance. When choosing the % to precalc, is build time linear (or nearly linear) to this? i.e. if selecting 10% going to be 3 times faster than selecting 30%? Also, is it fair to assume that if only 10% of the values are precalculated, that on average end user queries will have to hit 3 times as much data and therefore be about 3 times slower?
Sorry, doing this on a VM on a laptop, so testing various build configs is taking forever (I have yet to have a cube load actually finish). Guess I shouldn't be trying a 15 dim cube on a VM on a laptop, but trying to sell DBAs on the fact that it could enhance performance of our data mart.
Thanks,
Scott