[mercury-users] ICFP'00 Contest Results

Fergus Henderson fjh at cs.mu.OZ.AU
Wed Oct 4 12:52:02 AEDT 2000

On 04-Oct-2000, Tyson Dowd <trd at cs.mu.OZ.AU> wrote:
> On 03-Oct-2000, Ralph Becket <rbeck at microsoft.com> wrote:
> > Looking at the timings, I think we suffered badly for two reasons:
> > 1. Mercury's floating point performance sucks (both the OCaml and
> > Haskell compilers support unboxed floats); and
> Yep.  Improving floating point is a another project that we'd love
> someone to do, but it seems to be a reasonably big job.
> Recompiling in hlc.gc grade might help somewhat.  However there are a
> few unimplemented bits and pieces stopping us from running the ray
> tracer in this grade (compare is not implemented for univ yet).
> > 2. not having the bounding spheres optimization really shafted us
> > (at least, it looks like the judges were using the non-BS version).
> Yes, this seems to be the case.  IIRC we had some last minute bugs with
> that code, so took it out (correctness first).  

Right.  As it said in the README file:

        We implemented a bounding box space-partitioning algorithm,
	and it was about 50% faster.  Unfortunately it produced three
	pixels of acne in the shadows for a couple of the Tier-one
	examples, so we decided to submit our original version,
	which did not have such artifacts.

However, since the winning entry seems to have had some wierd
distortion effects, at least for the `golf' example, and yet won
anyway, I'm not sure if that was a good idea.  (Or are those
distortion effects just artifacts of the conversion to .jpg?)

Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at        |     -- the last words of T. S. Garp.
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe

More information about the users mailing list