From: Erik Enge
Subject: Lisp and 3D/mathematical computing.
Date:
Message-ID: <385567B5.41BCC4A8@src.no>
Hi!
Is there any good papers on Lisp and 3D computing lying
around, waiting to be read? I'm interrested in how Lisp
compare to C when computing mathematical formulas...
The best would probably be a 3D engine written in Lisp;
Common Lisp preferrably.
Thanks,
Erik Enge.
From: Robert Monfera
Subject: Re: Lisp and 3D/mathematical computing.
Date:
Message-ID: <385569D4.715BC0D9@fisec.com>
Erik Enge wrote:
>
> Hi!
>
> Is there any good papers on Lisp and 3D computing lying
> around, waiting to be read? I'm interrested in how Lisp
> compare to C when computing mathematical formulas...
Aren't the speed-critical parts of 3D engines implemented in hand-tuned
assembly?
Robert
From: Erik Enge
Subject: Re: Lisp and 3D/mathematical computing.
Date:
Message-ID: <38556FBC.703922A3@src.no>
Robert Monfera wrote:
>
> Aren't the speed-critical parts of 3D engines implemented in hand-tuned
> assembly?
Sure, but what about the rest? Is Lisp as fast as C? I guess the answer
is "implementation dependant", so the question should be: which CL
implementation to use when coding 3D engines?
Erik Enge.
From: Reini Urban
Subject: Re: Lisp and 3D/mathematical computing.
Date:
Message-ID: <3856bb0f.120464809@judy>
Erik Enge wrote:
>Is there any good papers on Lisp and 3D computing lying
>around, waiting to be read? I'm interrested in how Lisp
>compare to C when computing mathematical formulas...
>
>The best would probably be a 3D engine written in Lisp;
>Common Lisp preferrably.
Both do exist:
look for SIMLAB which contains also weyl.
http://www.cs.cornell.edu/Info/Projects/SimLab/
weyl has such a paper and simlab is one implementation in CL.
The docs don't do comparisons to C but weyl handles quite a lot of
algebra not seen elsewhere. Also some meshing functions.
There's also Genera but I don't have that unfortunately. There are some
more CL code samples on the net but that interesting as simlab.
--
Reini Urban
http://xarch.tu-graz.ac.at/autocad/news/faq/autolisp.html
In article <·················@src.no>,
Erik Enge <····@src.no> wrote:
>
> Hi!
>
> Is there any good papers on Lisp and 3D computing lying
> around, waiting to be read? [snip] The best would probably be a 3D
engine written in Lisp;
> Common Lisp preferrably.
>
Well, there's a 3D renderer written in /ANSI Common Lisp/ by Paul
Graham, chapter 9. It's not optimized for speed, but chapter 13 covers
that well.
Sincerely,
Douglas M. Auclair
Sent via Deja.com http://www.deja.com/
Before you buy.
>>>>> "DMA" == dauclair <········@hotmail.com> writes:
[...]
DMA> Well, there's a 3D renderer written in /ANSI Common Lisp/ by
DMA> Paul Graham, chapter 9. It's not optimized for speed, but
DMA> chapter 13 covers that well.
And that particular example was collectively optimized on this newgroup
some time ago. Dejanews should have the thread (floating point speed or
some such).
BM
Bulent Murtezaoglu <··@acm.org> writes:
> And that particular example was collectively optimized on this newgroup
> some time ago. Dejanews should have the thread (floating point speed or
> some such).
Here is a link to the thread of that discussion (if it is the one you
are thinking of):
http://www.deja.com/thread/332449950
Chris.
Erik Enge <····@src.no> writes:
> Hi!
>
> Is there any good papers on Lisp and 3D computing lying
> around, waiting to be read? I'm interrested in how Lisp
> compare to C when computing mathematical formulas...
C is nearly assembly language level. C will be faster but you will
pay the price in increased complexity, difficulty in debugging and
having to perform memory management manually.
lisp will offer a high level platform. sure, it may be slower
perhaps by a factor of two. still, O(N^k) algorithms stay O(N^k)
hence you will not see disasterous performance hits. if speed is
*really* important, you can call assembly routines from lisp. do this
after profiling.
lisp is probably more similar to matlab. lisp is interactive. you
enter your lisp environment and stay there rather than the C
edit/compile/test cycle. just like matlab, lisp does memory
management. lisp's memory management and language speed will beat
matlab. matlab's built-in matrix functions are written in C and will
likely beat a pure lisp implementation.
when algorithms are not static, i find lisp to be far superior to C.
for example, the romberg method of numerical integration involves
1) an interface taking a function (functions are first class objects
in lisp but not in C)
2) internally building a tableau from which the integral's value is
extrapolated. the number of tableau entries will vary according to
the particular intgral and are a priori unbounded. lisp's
automatic memory management is very handy here.
> The best would probably be a 3D engine written in Lisp;
> Common Lisp preferrably.
--
J o h a n K u l l s t a m
[········@ne.mediaone.net]
Don't Fear the Penguin!