Hi Luke,

There is/was an offline tool that would help with choosing subgoal
order, splitting, incrementalization, BDD library parameters, etc.
The way it works is by modifying the Datalog program or command-line
options, running the solver, and parsing the solve time.  If you are
interested in it, I might be able to dig it up.  Some version might be
checked into the source repository somewhere.  I'm not sure if it is
up-to-date with the latest bddbddb solver.  I also wanted to hook it
up to the active learning engine but never got around to it.

-John

On 6/14/07, Luke Simon < luke.simon <at> gmail.com> wrote:
> The PLDI tutorial mentions profile directed optimization of iteration order,
> yet doesn't mention how to use the feature.  Can bddbddb also do profile
> directed optimization of subgoal order, splitting, incrementalization, and
> other performance parameter directives?  I like the idea of writing a
> high-level analysis specification, and then running a profile based
> optimizer that tweaks every performance parameter: variable ordering,
> iteration order, subgoal order, splitting, incrementalization, etc.  So
> other than variable ordering, is there a feature for profile directed
> optimization of other performance parameters, and if so, how are they used?
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> bddbddb-devel mailing list
> bddbddb-devel <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bddbddb-devel
>
>