POV-Ray : Newsgroups : povray.beta-test : radiosity in 3.5 - any improvments towads parallelization? Server Time
29 Jul 2024 18:29:45 EDT (-0400)
  radiosity in 3.5 - any improvments towads parallelization? (Message 1 to 6 of 6)  
From: George Pantazopoulos
Subject: radiosity in 3.5 - any improvments towads parallelization?
Date: 6 Apr 2002 16:07:09
Message: <3caf637d$1@news.povray.org>
Is the 3.5 team doing anything about the radiosity problem that occurs when
multiple computers are assigned to render pieces of the same frame? See Web
page here: http://www.wozzeck.net/images/pmp/ It would be nice to have an
official solution to the problem, even though it seems not to be a simple
problem. Any comments welcome.

Thanks,
George Pantazopoulos


Post a reply to this message

From: Jan Walzer
Subject: Re: radiosity in 3.5 - any improvments towads parallelization?
Date: 6 Apr 2002 16:11:54
Message: <3caf649a@news.povray.org>
No !

There are workarounds, but there's no parallelization of
Radiosity in POV3.5

There will be none until POV4 !


Post a reply to this message

From: Thorsten Froehlich
Subject: Re: radiosity in 3.5 - any improvments towads parallelization?
Date: 6 Apr 2002 16:14:00
Message: <3caf6518@news.povray.org>
In article <3caf637d$1@news.povray.org> , "George Pantazopoulos" 
<the### [at] attbicom> wrote:

> Is the 3.5 team doing anything about the radiosity problem that occurs when
> multiple computers are assigned to render pieces of the same frame?

POV-Ray does not know you are going to use multiple parts and assemble them
later, so how can it do anything about this?  The problem cannot really be
solved without those computers communicating with each other.  And as you
know, network rendering is not a feature of POV-Ray 3.5.  So the answer to
your question is a simple: Nothing.

    Thorsten

____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg

I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org


Post a reply to this message

From: George Pantazopoulos
Subject: Re: radiosity in 3.5 - any improvments towads parallelization?
Date: 6 Apr 2002 17:00:58
Message: <3caf701a$1@news.povray.org>
Thank you for answering, Thorsten
    I asked this question because I have been comtemplating writing my own
program to control multiple machines running POV-Ray, until I stumbled onto
the radiosity problem. I've read the overview in the docs of how 3.5 handles
radiosity, but can you show me where to read about this in much more depth?
Is it possible to get POVRay to output octree data?

Regards,
George Pantazopoulos

"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3caf6518@news.povray.org...
> In article <3caf637d$1@news.povray.org> , "George Pantazopoulos"
> <the### [at] attbicom> wrote:
>
> > Is the 3.5 team doing anything about the radiosity problem that occurs
when
> > multiple computers are assigned to render pieces of the same frame?
>
> POV-Ray does not know you are going to use multiple parts and assemble
them
> later, so how can it do anything about this?  The problem cannot really be
> solved without those computers communicating with each other.  And as you
> know, network rendering is not a feature of POV-Ray 3.5.  So the answer to
> your question is a simple: Nothing.
>
>     Thorsten
>
> ____________________________________________________
> Thorsten Froehlich
> e-mail: mac### [at] povrayorg
>
> I am a member of the POV-Ray Team.
> Visit POV-Ray on the web: http://mac.povray.org


Post a reply to this message

From: Christoph Hormann
Subject: Re: radiosity in 3.5 - any improvments towads parallelization?
Date: 7 Apr 2002 04:23:25
Message: <3CB001FD.15358ED2@gmx.de>
George Pantazopoulos wrote:
> 
> Is the 3.5 team doing anything about the radiosity problem that occurs when
> multiple computers are assigned to render pieces of the same frame? See Web
> page here: http://www.wozzeck.net/images/pmp/ It would be nice to have an
> official solution to the problem, even though it seems not to be a simple
> problem. Any comments welcome.

There's a very commonly used method to render radiosity scenes in stripes
developed by Gilles Tran.  You just set pretrace_start and pretrace_end to
1 and only use the samples calculated during final trace.  This works very
well with low error_bound, but there could be problems and speed
disadvantages with higher values.  Rendering the different stripes on
different machines should not be a problem.  

Another possibility is to calculate all radiosity samples in advance, save
them and using this data for all the machines rendering the scene.

Christoph

-- 
POV-Ray tutorials, IsoWood include,                 
TransSkin and more: http://www.tu-bs.de/~y0013390/  
Last updated 18 Mar. 2002 _____./\/^>_*_<^\/\.______


Post a reply to this message

From: Thorsten Froehlich
Subject: Re: radiosity in 3.5 - any improvments towads parallelization?
Date: 7 Apr 2002 05:09:29
Message: <3cb00cc9@news.povray.org>
In article <3caf701a$1@news.povray.org> , "George Pantazopoulos" 
<the### [at] attbicom> wrote:

> Thank you for answering, Thorsten
>     I asked this question because I have been comtemplating writing my own
> program to control multiple machines running POV-Ray, until I stumbled onto
> the radiosity problem. I've read the overview in the docs of how 3.5 handles
> radiosity, but can you show me where to read about this in much more depth?
> Is it possible to get POVRay to output octree data?

Yes, it is.  It has the extension ".rca" and contains a dump of the octree use
when continuing a trace.  You can check it is the 3.1 source code for now, in
radiosity.c functions called ot_read_file and ot_save_tree are used (you find
them in octree.c).  In 3.5 you can also access the reading and writing
directly, see section "6.11.11.2.15 Save and load radiosity data".  However, I
have never really checked how much space this data will take.  If you render
the scene at a lower resolution and then transfer this file to all computers
to to the final render you may be able to work around the lack of a shared
octree and still get a half way reasonable radiosity, but it will most likely
not help a lot (as far as speed is concerned) as you have to generate this
file on a single computer first....

    Thorsten

____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde

Visit POV-Ray on the web: http://mac.povray.org


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.