Thanks for all the recommendations. But it still worries me that web2py (or uwsgi) does not automatically (a) cache/reuse unmodified code and (b) release memory after requests are served (or while idle), after each web request, and regardless of how small or big models actually are.
Carlos
