Great, the issue seems to be fixed as the spreadsheet is generating fine
with the latest nightly build. I'm now going to profile and compare the
2.5.1 and 3.0 versions with my app that generates a decent size spreadsheet
( 6000 rows, 15 columns).

Any insight on what performance improvements were made from 2.5.1 to 3.0?
Can we expect a lower memory footprint or any leak fixes? To me CPU time
during excel generation is not as important as the memory footprint. As I
explained in a previous mail, we need to be able to handle several
concurrent downloads of generated spreadsheets ( 6K rows) and if the memory
footprint of each generation is ~ 300 MB (which is close to what I was
observing) , this limits the number of concurrent downloads to 4 or 5 which
isnt very scalable.

I have added queued execution of excel generation to control the maximum
number of concurrent downloads, but I still would like to see a lower memory
footprint / leak fixes with 3.0.

Thanks,
Sanjiv


On 1/12/07, Nick Burch <[EMAIL PROTECTED]> wrote:

On Fri, 12 Jan 2007, Sanjiv Jivan wrote:
> Do you think the fix was made after Dec 12th, 06?

I can't remember the exact date, but it might well be after then

> If so, are nightly builds available anywhere or would I need to build
> off SVN?

http://encore.torchbox.com/poi-svn-build/

Nick

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
Mailing List:     http://jakarta.apache.org/site/mail2.html#poi
The Apache Jakarta Poi Project:  http://jakarta.apache.org/poi/


Reply via email to