Recalculation of Data

[ParisHilton]

Hi forum,
we use our cubes for BI reporting, typically loaded with data overnight and queried through the day, with no new data going in.

What I am finding (looking at process monitor) is that if I run a query (Via BI report) there is a lot of processor usage. This I can understand the first time I run a report. But if I run the same report again (no parameter changes), I'd expect my cube to get it's results from RAM and barely touch the CPU.

Might I have set up something by accident that is forcing recalculation? I've got skipcheck and feeders in there, not that I suspect that would cause a problem

Cheers
ps - I doubt it makes any difference, but I recently reorderd dimensions, putting my measures dimension first, which cut the cube to about a quarter of it's size and drastically improved performance. Fairly sure this is unrelated, but possibly points to corruption or something nasty!

[stephen waters]

typically loaded with data overnight and queried through the day, with no new data going in.

Couple fo questions,
- You say "typically" loaded overnight. Do you have ANY changes going on during the day because this will cause the complete calculation cache of input and dependent cubes to be flushed
- When you say you are running the same BI report, is it accessing the same data within TM1 or is it the same format BI report ( eg P&L) but looking at different costs centres\regions?

[ParisHilton]

Poorly written post, there are two loads that happen during the day. But my testing occurs after both these loads have happened, so data is consistent.
It's EXACTLY the same BI report returning the same data.

I like your troubleshooting style.

[Catherine]

Try to play with VMM parameter in the }CubeProperties cube.
I once had the same problem as you and it solved it.