I am currently in a TM1 project and have some questions about performance tuning using FEEDERS. Below is my question, hope you can kindly give me some hints/advice. Thank you.
- 1. Suppose that the reference between cubes A, B, and C is A > B > C, i.e. cube C uses a DB() function to reference cube B and cube B references cube A. I added a FEEDERS in rules of cube B and it shows the cells in cube C are fed when I uses Trace Feeders on cube B. However, the consolidation on cube C is still missing the values obtained from cube B. I found some discussions in TM1 forums said FEEDERS should also be added in rules of cube A. Is it correct?
- 2. Cube C has 9 dimensions and 4 dimensions have large number of elements (Organization, Product, G/L Accounts, etc.). After adding FEEDERS on cube B, the used memory space keeps increasing when I opened cube C and finally I got an OutOfMemory exception in system log. Does it mean I over-feed cube C?
- 3. If cube B has fewer dimensions than cube C, how should I feed cube C since some dimensions in cube C are not in cube B? Currently, I add an element as top consolidation element in such dimensions, which leads to over-feeding. Is there any better ways to feed?