I would just like to know if it would be inefficient if I created a hundred or more view streams or view accessors from a Memory Mapped File object per process, and I had a few or more processes running that shared this file. (By file I mean memory-mapped file that is not persisted to the disk.) It is more efficient to create and destroy one view stream / view accessors many times? Or do I need to rework my implementation of what I'm working on so that it depends on only one or a few views accessors / streams?

Any help would be appreciated, and this is urgent because I am trying to get something done as soon as possible.