I agree with Rolf who disagrees with this claim :-)
> One goal of one-sided is to allow more natural programming of
> Typically they have a data area (the window) where the other processes
> can put (or get) their results.
> Looking at applications with irregular grids and with automatic
> load balance, they dont want for each process an own data region.
Exactly. In chemistry apps this would correspond to some large ( >100MB in size)
distributed matrix with processes accessing dynamically different subsections
(around 1KB in size). There is no predefined static access pattern, and therefore
mapping these small subsections into separate windows to avoid concurrent
accesses is not practical. Also locking is expected to be an unnecessary and
heavyweight operation that would kill performance.
> Therefore it is important to allow one window with many non
> overlapping regions which dynamically change in position and
> length channging.