From: janburse@fastmail.fm   
      
   We recently presented a fast constraint solver   
   termed Railgun CLP(FD) that modelled attributed   
   variables simply via ‘$ATTR’/2 compounds and could   
   deal with integer dif/2 constraints. In this   
   instalment we allow (#\=)2 constraints and demonstrate   
   that it can be ported to SWI-Prolog.   
      
   Using polyfill for ‘$SEQ’/2 from Dogelog Player,   
   we observed that running Railgun CLP(FD) inside   
   SWI-Prolog gives a 2–3x speed-up for the Queens   
   example, on both 32-bit and 64-bit. On the other   
   hand the price tag for big integer flexibility   
   seems to be a factor 35x slow down.   
      
   See also:   
      
   Porting Railgun CLP(FD) to SWI-Prolog   
   https://medium.com/2989/e9f2ef4e6878   
      
   Mild Shock schrieb:   
   >   
   > Hi,   
   >   
   > Many existing and evolving constraint logic   
   > programming projects resemble some ancient   
   > invention of gunpowder. For example SWI-Prologs   
   > 9.3.35 corouting for delayed goals is mainly   
   > based on unify hooks. We show how verify hooks,   
   > already used in formerly Jekejeke Prolog, can   
   > be braught to Dogelog Player in a 100% Prolog fashion.   
   >   
   > Since the recent version of Dogelog Player   
   > supports cyclic terms, we could let the Jini   
   > out of the bottle, and provide the experimental   
   > library(edge/railgun) to model delayed goals with   
   > nothing else than Alain Colmerauers rational trees.   
   > The result is a Lean CLP of ca. 100 lines of code,   
   > that already provides a simple constraint (#\=)/2   
   > and a global constraint all_different/1.   
   >   
   > The results are encouraging. For problems that   
   > are not over constrained, Dogelog Player leaves   
   > existing Prolog systems clearly behind, showing   
   > a 2-3x times speed-up against SWI-Prolog and a   
   > 20-30x times speed-up against Trealla Prolog.   
   > For more constrained problems we suggest ommiting   
   > forward checking in favor of a form of ahead of   
   > time (AOT) variable ordering. With this approach   
   > and for Sudoku problems we are then in the midfield   
   > between SWI-Prolog and Trealla Prolog.   
   >   
   > Bye   
   >   
   > See also:   
   >   
   > Lean CLP for Dogelog Player   
   > https://qiita.com/j4n_bur53/items/addf1fc86856dd682dcb   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)