Porting the SWI-Prolog benchmark suite: comparing 8 Prolog systems

Nice to know
, ,
31

Thanks for the input. So, here is take 2:

  • Updated the “dummy” loop timing as contributed by @jfmc
  • Fixed queens_clpfd.pl. The old code posted the constraints, but did not solve them. There is something weird going on in SICStus here. SICStus is surely much faster in clp(fd) in general. I asked for a response, so this may get updated with either an explanation or a fix.
  • Some driver updates, notably an option to regenerate the charts in different combinations without rerunning the benchmarks.

Some classic systems

classics

  • Updated Ciao Prolog to 1.22.0
  • Updated Ciao chart to include @jfmc’s improved loop.

The “modern” systems vs. SWI-Prolog

modern

  • Updated Scryer Prolog to 0.9.3. They made serious progress. Congrats! The queens_clpfd.pl and the sieve.pl benchmarks have been added. The ISO predicates number/1 and retractall/1 have been added. I had to made more changes to to get the code loaded. Creating a module with the programs and some support predicates somehow did not work anymore (predicates became invisible). Loading a file programs.pl from directory holding a subdirectory programs silently loaded nothing until I added the .pl suffix. The sieve bar is cut at 20, but the actual value is 359.

I will add a pointer to these results in the original post.

2 Likes