what I found amazing, is when the user go at the lower detail!
top and bottom levels provides a excellent performance.
I can drill down in seconds while with AS2000 I have to wait more (or simply
I disallow the use some dimensions to insure that the performance was good)
the server cache really improve the performance in my case because I have
some scope calculations; the cell is evaluated 1 time and its all!!! AS keep
the result in memory.
I found that AS2005 consume more memory, but my reaction is more like: well
now AS has a real cache! :-)
the compression also help us; I can access my cube other a slow network and
I can retrieve a lot of detail in seconds instead of minutes (like crossing
the 2 biggest dimensions at the lowest level takes only 2 to 5 seconds)
so, yes, its impressive the new perofrmance.
in my case I talk about test cubes for a total of 40 millions of rows and 2
dimensions with 10 000 members on each one with a lot of attributes.
"Jesse O" <jesperzz (AT) hotmail (DOT) com> wrote
Quote:
I'm getting 7-10x better performance on a lesser box and I'm having
trouble believing it.
Numbers are matching up...I keep double checking it, just can't believe
it.
Has anyone else experienced this kind of performance gain? |