That's not a plugin issue, but yes, region render takes just as much memory as a full render...it just renders that part faster. However, with the 1.6.8 plugin, you should rarely need to use region render. If you use the Camera in 'Manual' Size Mode for your image output, you can then use the command 'Maxwell_PickFilmSizeRectangle' to define a region you'd like to render - it will have a perspective match with the image size you've defined using the Camera's film-size. Here's an example:
1. a viewport, set up to render a certian size of film, at a random output resolution:
2. the same viewport, after using Maxwell_PickFilmSizeRectangle to define a new region:
3. combining the images together:
The only restriction is that the region picked can't be more than 100% of the film width/height off-center, since this uses Maxwell's Shift Lens feature to define the rendered frame. Maxwell_PickFilmSizeRectangle Rectangle is in the Camera toolbar, third button from the left. If you work this way (i.e. not using 'Viewport' for the Camera's Size Mode), your output resolution will not be linked to the size of the Rhino viewport, so it won't get screwed up when you resize your viewports or move toolbars around.
I hope it helps,
JD