As the performance improvement due to unboxing given by Juho Snellman's Lisp
macro in his implementation of my ray tracer rivals the algorithmic
improvements that I made and it is applicable to other languages, I think
I'll include such unboxing as a final optimisation, giving 5 programs in
each language.
Therefore, SBCL will perform much better relative to the other languages for
this final implementation. Moreover, Juho's macro is sufficiently succinct
and reusable that the verbosity may also be relatively better.
I'm wondering if people have any other ideas for improvements to this
benchmark that would show Lisp in a more positive light?
--
Dr Jon D Harrop, Flying Frog Consultancy
http://www.ffconsultancy.com