Performance concerns when downloading reports/large amounts

[Radhika]

Hi All,
We are supporting a TM1 environment: Windows 2003, Standalone, TM1 Architect, TM1 Perspectives & TM1 Web. 8 GB RAM (32-bit)
We are accessing TM1 web reports but it is taking lot of time while downloading the reports. We tried to increase the RAM size but that's not working. Rules are optimized propely ( we do not have any rules affecting the consolidated elements).

Could please let me know what would be the critical areas that needs to be looked in to increase the downloading speed.

Thanks,
Radhika

[tomok]

We tried to increase the RAM size but that's not working.

No surprise there since in a 32-bit Windows environment no application can theoretically access more then 2GB of RAM and in most systems, due to Windows overhead, it's really more like about 1.5GB. There is an internal Windows switch you can set to make that limit 3GB (which really results in about 2.5GB for TM1) but that's it. Have you checked Peformance Monitor and the TM1 system performance cubes to see how much memory your application is really using? If you are at the limit you could be switching to paging files which kills performance. Have you tried the same reports in the Excel and Perspectives to see if it's only the Web that is slow?

Rules are optimized propely

Don't mean to be argumentative here but exactly how would you know that? If you don't have enough experience to know how to troubleshoot performance issues in TM1 how can you know enough about rules, feeders, etc., in order to make such a bold statement? I'm not saying they aren't optimized, just don't take that option off the table.

[jim wood]

I agree with Tomok. The only way to truely know that you have all rules optimised is to have non at all. Rules and feeders especially can be a lot of the time different in terms of optimisation based on the circumstances, and these may change over time as the data changes.

[lotsaram]

Quite aside from whether the model itself is optimized or not (I also suspect not) and whether the hardware is underpowered or adequate what about the reports themselves ... How large are they? Do they contain thousands of rows? If they are websheets as opposed to views how large and complex is the underlying Excel file? etc., etc. Fact is if you have a "report" of dozens to hundreds of columns and thousands of rows that is better described as a data dump as opposed to a report then it is going to take quite some minutes to download to Excel no matter what you do.

[Radhika]

Thank you all for your suggestions.