#Speed
4 messages in this thread
Hi Ed,
I hope you use POE libs for calculating your chaos and collisions in AniPro.
If not you should do this, because POE libs run 3 times faster as the orginal
POCO code.
Ingo
<Hi Ed,
I hope you use POE libs for calculating your chaos and collisions in AniPro.
If not you should do this, because POE libs run 3 times faster as the orginal
POCO code.>
Hi Ingo. I haven't used POE yet. Do you mean that there are libraries already
available for some of these calculations?
Right now I am experimenting a lot, and modifying the code after each run.
Would I still want to use POE in that case? If you don't mind, I would
appreciate hearing more details on what's involved.
Here are more details on what I am doing:
I got started by looking through my books on how to calculate the motion, under
gravity, for several objects, and couldn't find anything. So, I made a few
back of the envelope manipulations (starting with 1/2*g*t*t) and tried what I
got. The results were so interesting that I haven't had the time to recheck the
equations, and assumptions, I am using. Do you know where I could find a
description of how to calculate these trajectories?
By the way, I'll send my programs to anyone who would like to try them.
The program I am working on right now, calculates trajectories for two galaxies
whose mass is concentrated at one point. One of the galaxies has about 180
stars. The stars are influenced by the two masses, but generate no gravity
field themselves. I'm trying to reproduce the results in "Violent Tides Between
Galaxies" from Scentific American of 12/73.
They show that close approaches can generate galaxies with spiral arms. The
problem I am having, is that as soon as I start one galaxy, with stars in
random positions, all rotating in the same direction and at speeds that result
in stable circular orbits, the darn thing develops spiral arms ON ITS OWN
(before the other galaxy approaches it).
Let me know what you think.
Ed K
Ed,
I am not the man for math problems in galaxies, that doesn't mean that Iam not
intressted in it. I read a lot of books about the begining universe, micro/mkro
kosmos, but without any aspct of math. Perhaps you sould ask the NASA or a
astronomic institute for more infos. I know that the collision of galaxies was
already calculated and
visualisted in super computers.
Back to the POE'S. POCO is an interpreter, that means your POCO (C) code will
be translated into CPU instructions during runtime. You lost a lot of CPU power
for this translations but you have the atvantage manipulate your code
suddendly. A POE lib is nearly the same POCO code, but it's translated before
starting the main POCO program, so you can use the full CPU speed for extensive
calculation like chaos algorithems or particle systems. To create your own
POE's you need a C-Complier like
METAWARE HIGH C and the PHARLAP DOS EXTENDER/LINKER. After compiling your code
you can use it as a C function in POCO. Because the time intensive code is
already compiled you run in full CPU speed. I tested this with the mandelbrot
algorithem which run 3 times faster with POE precompiling. OK, that should be
enougth for the moment.
BTW, ANIMATOR's POCO is a great tool to program some small utilities into
Animator. For grater Animator independent software, you should use a REAL C
Compiler with a userinterface and an EDITOR!!, debugger, ……………….
This is more productive for your work. When you have finished you can esaliy
import the main routines into POCO and use the atvantage of image/Filc support.
Ask your dealer for the AUTODESK FLIClib for support outside Animator and the
POE KIT. Welcome in the wonderful world of software engeniering 🙂
I know Animators POCO interpreter is easiy to handle compared with a real
compiler, but belive me, after 500 lines source code a real compiler is better.
(My greates POCO has 1500 lines, then I loose the overview<g>).
If you like, and I have some time, I can compile the time intensive routines of
your POCO into a POE.
Ingo
Sorry for the delay in replying. I haven't done much else
since I figured out how to use AniPro/POCO to play with Chaos and
Galaxies.
<to create your own POE's you need a C-Compiler like METAWARE HIGH C
and the PHARLAP DOS EXTENDER/LINKER.>
You talked me into it! But the price is a shocker! I called
Pharlap today to order the linker and found that it is part of a
$500 SDK that includes a C compiler. I already have the Watcom C
compiler (bought it to try my hand at IPAS routines, but never got to
it).
Is there a cheaper way to get all I need (without paying for
another C-compiler)?. Whatever I get should work for IPAS also.
The galaxy stuff might be headed that way.
<I am not the man for math problems in galaxies>
I finally took the time to figure it out, and I have it all
working now.
<If you like, and I have some time, I can compile the time intensive
routines of your POCO into a POE.>
Thanks. I may take you up on the offer.
But I have a problem. I modify many parameters like galaxy/star
positions, velocity, colors, etc. by simply editing the POCO code.
If I compile the code, I would need an alternate way to edit
parameter values. POCO only seems to provide a way for the user to
enter one value at a time. Any thoughts?
Ed K