[PPL-devel] Polyhedra/test Makefile.am Makefile.in bug3.cc ...

[Enea Zaffanella]

> Log message:
> 	bug3 shows what may or may be not a bug.
> 	In any case, Polyhedra::conversion() is entered with a big
> 	(but not terrible) polyhedron and control seems not to comes back,
> 	while memory usage seems to grow slowly but continuously.
> 	It may be "the concrete effects of combinatorial explosion"
> 	or a genuine bug.

I have started playing with the example.
When considering the first 33 constraints, it returns almost immediately
that the polyhedron is not empy;
when considering 34 constraints, it returns after some time (say, a couple
of minutes on gandalf, which is slow and has few memory);
with 35 constraints it is still running after 6 minutes of CPU time.

It would be also interesting to experiment with different segments 
of the matrix, e.g. by using the LAST rows instead of the first ones ...

Enea.