[Date Index] [Thread Index] [Author Index]

Re: Mathematica and F#

• To: mathgroup at smc.vnet.net
• Subject: [mg91629] Re: Mathematica and F#
• From: Jon Harrop <jon at ffconsultancy.com>
• Date: Sat, 30 Aug 2008 01:50:26 -0400 (EDT)
• References: <g93b2d$km1$1@smc.vnet.net> <g95k7v$53r$1@smc.vnet.net> <g962t3$c9c$1@smc.vnet.net>

rych wrote:
> Could someone please post a test program coded both in F# and
> Mathematica where F# outperforms Mathematica?

The ray tracer benchmark described here is one such program:

  http://www.ffconsultancy.com/languages/ray_tracer/

The hierarchical spherical bounding volumes used in this ray tracer are
comparable to a wide variety of hierarchical algorithms and data structures
commonly seen in scientific computing, e.g. the Fast Multipole Method used
in cosmology and molecular dynamics.

I have uploaded a Mathematica 6 implementation of this benchmark here:

  http://www.ffconsultancy.com/languages/ray_tracer/code/1/ray.nb

The OCaml code is currently the most comparable to F# and is over two orders
of magnitude faster than the Mathematica. I'll upload some real F# ASAP...

-- 
Dr Jon D Harrop, Flying Frog Consultancy
http://www.ffconsultancy.com/products/?u