3 ms·

> The downside was that it took several minutes to compile <100 line sources that produce <1000 instruction programs. This sounds like the Stalin Scheme compil

by derleth 13y ago

> The downside was that it took several minutes to compile <100 line sources that produce <1000 instruction programs. This sounds like the Stalin Scheme compiler. Simple benchmark (vs C (gcc), Stalin wins): http://justindomke.wordpress.com/2009/02/23/the-stalin-compiler/ http://justindomke.wordpress.com/2009/02/23/the-stalin-compi... GUI addon: http://code.google.com/p/stalingui/ http://code.google.com/p/stalingui/ http://community.schemewiki.org/?Stalin http://community.schemewiki.org/?Stalin Download link: https://engineering.purdue.edu/~qobi/software.html https://engineering.purdue.edu/~qobi/software.html

ANTSANTS 13y ago
Stalin is very impressive, but > Simple benchmark (vs C (gcc), Stalin wins) Did you read the comments? > Nice, but part of the reason the C version is slow is that you are using abs instead of fabs. abs takes an int as an argument (this should be a compiler warning!) – passing a double other than zero always returns a large number. The C version is basically iterating till the error is below double’s resolution. > Excellent point! Making the switch to fabs brings the gcc-inline speed down almost exactly to what Stalin is getting.