Optimise my ray tracer - comp.lang.java.programmer

Message from discussion Optimise my ray tracer

Skip

View profile

More options 2 Jun 2005, 15:15

Newsgroups: comp.lang.java.programmer

From: "Skip" <a...@b.invalid>

Date: Thu, 2 Jun 2005 16:15:36 +0200

Local: Thurs 2 Jun 2005 15:15

Subject: Re: Optimise my ray tracer

  Reply to author
  |
  Forward | Print | View thread
  | Show original
  | Report this message | Find messages by this author
  

"Jon Harrop" <use...@jdh30.plus.com> wrote in message

news:429e5795$0$7560$ed2619ec@ptn-nntp-reader03.plus.net...

> I've written a mini ray tracer for the computer language shootout. The
> original version was written in OCaml which I ported to C++:

> http://www.ffconsultancy.com/free/ray_tracer/comparison.html

> I have since ported the program to several other languages, including
Java.
> Currently, the Mlton-compiled SML, OCaml, C++ and Fortran are the fastest,
> and Java trails a long way behind. I am concerned that this is because I
am
> a much better OCaml/C++ than Java programmer so I'm asking for advice
here.

The first run of this code is done in interprete-mode. If you run it a few
times, things get faster. Here is my benchmark (with n=32)

iteration #0: 474.3ms
iteration #1: 198.0ms
iteration #2: 151.1ms
iteration #3: 154.9ms
iteration #4: 153.1ms
iteration #5: 155.9ms
iteration #6: 149.5ms
iteration #7: 149.3ms

P4 2.4GHz (running at 1.8GHz)

What kinds of differences do you see for n=32 in calc-duration, comparing
Java and C++?

Reply to author Forward