leakage or no leakage

[dizzy]

Hi,
i have an app that shows a foot print in task manager of around 150K when it starts. Over the next 10 to 20 mins or so it climbs to about 500K and then drops back down to 250k. this then repeats..it climbs to ~500k and then drops back to 250k.
It doesn't seem like a memory leak since gc seems to be happening occassionally to bring the foot print back down, but i'm trying to undersand what may be going on here to cause the signifant increase over time and why it takes so long for gc to clean up.

After running ant and take some snapshots i see that between snapshots, there is not that much of a difference in size(initial size is about 47.1MB and after the 2nd snapshot, its about 47.8MB). I do see large number of records in gen2 from the get go however (about 47.5mb).

any advice on what may be happening here and how to understand better?

[Chris.Allen]

All I can recommend is that you go through all our help on the memory profiler and this should point out some candidates for perhaps an overuse of the gc. Maybe you can refactor yor code such that objects don't hang around for any longer than they really need to (avoiding a generational problem). More than that, I really cannot say. We (red gate support) can, if we have enough time, take a look at profile results and analyse but please make sure it contains lots of data (snapshots).