Thank you Peter,
Indeed, i am very aware of the rules of optimization, and am therefore spending most time to complete functionality, yet as i continue to add functionality I see a noticeable slow down.
One key uncertainty that comes with it is whether I will be able to increase performance significantly in prolog or whether I will have to rewrite some core in C/C++ (or Rust).
Another uncertainty is whether Prolog is already optimized enough so that a move to, say, C, will not make a significant difference – an outcome like this would need deep rethinking what to do – including considering hardware acceleration already at this earlier stages.
Finally, i may have a deadline in 3 months where I will need to have resolved these questions in principle and possible in practice, to some extend, as well.
I am therefore trying to analyze in parallel, at least in principle, and with a bit of tinkering, where improvements can be had.
Dan