That is far easier for me to understand. Thank you. I hope other readers will now understand your questions and be able to answer them. I also understand why you cannot use the session extension. I do have one caution for you. You can test several configurations by making a lot of changes as fast as possible. And you will get a different speed from each one and find out which is fastest. Good. However, this reflects only the situation where you are making a lot of changes as fast as possible. If you test a situation where the changes are happening less often, a different configuration can be faster. So your result may be correct, but not useful for any real situation. Some reasons are that caching happens at several levels in your computer and storage subsystem, and some systems wait for inactivity, or for the cache to be full, before writing their cache. And your computer tries to write the cache at the same time as running operations in other processes. Things happen with different interlacing, in a different order. So if your real application is not going to make a lot of changes as fast as possible, this may not be such a useful test.