Our application is very large and processor intensive. The programmatic API to take snapshots is something that we think could be very useful. I've been asked to find out what if anything can be done to reduce the overhead of taking snapshots. For example, if a snapshot is manually triggered then not only is there effort put into taking the snapshot in the first place but also some effort goes into analyzing the results. In our case, we'd just like to take the snap but delay doing any analyzation until a later time. Is this possible? We'd also like to investigate to what degree is it possible to reduce the overhead involved in having the memory profiler active. My initial thoughts are that you can't take a snapshot without having the memory profiler active so there is little if anything that can be done to reduce the overhead involved in taking snapshots. Is this an accurate representation of the facts?
Richard Lewis Haggard