 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hello all,
For the past several months, in my spare time, I've been developing
parallel rendering software for the POV-Ray community. I am bending over
backwards to make it artist friendly. It supports 48-bit PNG output, will
have special handling for radiosity (I have some tricks up my sleeve and
have had encouraging results), and it allows any mix of Windows and/or Unix
machines to be used as render nodes. This all works with vanilla POV-Ray
3.5 - no special version is required.
Right now the multithreaded Master program runs on Windows and it is
targeted at artists who want to render high quality still shots.While my
program is still in the prototype stage, I've reached the point where I'm
doing distributed renders at home on my two machines, and sometimes also on
a third machine over the internet.
What I'm humbly asking you for is some high-quality test scenes. There
were several good ones I've found already in p.b.s-f, but can anyone suggest
or submit other production quality scenes? In particular, I'm looking for
realistic-looking radiosity scenes that go far beyond the cornell box,
and/or media scenes. The longer rendering time, the better :)
Thanks, suggestions welcome
George Pantazopoulos
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
George Pantazopoulos wrote:
> What I'm humbly asking you for is some high-quality test scenes. There
> were several good ones I've found already in p.b.s-f, but can anyone suggest
> or submit other production quality scenes? In particular, I'm looking for
> realistic-looking radiosity scenes that go far beyond the cornell box,
> and/or media scenes. The longer rendering time, the better :)
Have you tried benchmark.pov from the distribution?
--
Ken Tyler
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hey Ken,
I sure have, but now that you mention it, I'm going to try it again :)
Thanks,
George
>
> Have you tried benchmark.pov from the distribution?
>
> --
> Ken Tyler
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
George Pantazopoulos wrote:
>
> Hello all,
> For the past several months, in my spare time, I've been developing
> parallel rendering software for the POV-Ray community. I am bending over
> backwards to make it artist friendly. It supports 48-bit PNG output, will
> have special handling for radiosity (I have some tricks up my sleeve and
> have had encouraging results), and it allows any mix of Windows and/or Unix
> machines to be used as render nodes. This all works with vanilla POV-Ray
> 3.5 - no special version is required.
>
> Right now the multithreaded Master program runs on Windows and it is
> targeted at artists who want to render high quality still shots.While my
> program is still in the prototype stage, I've reached the point where I'm
> doing distributed renders at home on my two machines, and sometimes also on
> a third machine over the internet.
Sounds good. It would be great if the master program would run on several
platforms of course. I know this is probably more work but when you
distribute a long render you will want to run the master on a stable
machine that can run permanently - and this is often not a windows
machine...
> What I'm humbly asking you for is some high-quality test scenes. There
> were several good ones I've found already in p.b.s-f, but can anyone suggest
> or submit other production quality scenes? In particular, I'm looking for
> realistic-looking radiosity scenes that go far beyond the cornell box,
> and/or media scenes. The longer rendering time, the better :)
You should try the benchmark and balcony scenes in the 'advanced'
directory. The 'advanced/blocks/stackerday.pov' scene might also be good
for testing handling of radiosity although it does not use pretrace which
is somewhat atypical for radiosity scenes.
It would particularly be important to take care that it works correctly
for large size renders.
Christoph
--
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 28 Feb. 2003 _____./\/^>_*_<^\/\.______
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
noticias 3eabf2f9$1@news.povray.org...
> What I'm humbly asking you for is some high-quality test scenes. There
> were several good ones I've found already in p.b.s-f, but can anyone
suggest
> or submit other production quality scenes? In particular, I'm looking for
> realistic-looking radiosity scenes that go far beyond the cornell box,
> and/or media scenes.
Have you tried the zips sent to the IRTC?.
I suppose you have a lot of scenes there
> The longer rendering time, the better :)
I think the winner image from the Loneliness
round (made by Gena) could satisfy your needs ;-).
Bye
Txemi Jendrix
http://www.txemijendrix.com
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> The longer rendering time, the better :)
Do you want long image rendering times, or are long parsing times good as
well?
Fingolfin
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi Fingolfin,
Long parsing times are a problem when doing tiled parallel renders.
Currently povray loads, parses, renders, and quits each and every time it
renders a tile. If the parsing time as a significant fraction of the total
time, it would negate much of the speed gained. To solve this would probably
require patching povray so that it can parse just once, stay open, and
render tiles when it recieves messages from the Master (probably via
stdin?). I will eventually have to deal with long parsing times, so if you
know of any good scenes, I'll take em.
Thanks,
George
"Fingolfin" <til### [at] hotmail com> wrote in message
news:3ead98ae@news.povray.org...
>
> > The longer rendering time, the better :)
>
> Do you want long image rendering times, or are long parsing times good as
> well?
> Fingolfin
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> Sounds good. It would be great if the master program would run on several
> platforms of course. I know this is probably more work but when you
> distribute a long render you will want to run the master on a stable
> machine that can run permanently - and this is often not a windows
> machine...
>
I agree :) I would like to have it run on other platforms. Much of my
program is generically coded in C++, and should be portable. However, the
major stumbling block to portability is with regard to threading and
processes. Win32 has no native implementation of pthreads. However, I
recently found this library. http://sources.redhat.com/pthreads-win32/ Have
you heard anything good about it?
Also, I'm not sure what a platform independant way to start external
programs as processes would be. _execve() ?
>
> You should try the benchmark and balcony scenes in the 'advanced'
> directory. The 'advanced/blocks/stackerday.pov' scene might also be good
> for testing handling of radiosity although it does not use pretrace which
> is somewhat atypical for radiosity scenes.
>
> It would particularly be important to take care that it works correctly
> for large size renders.
>
>
What are your opinions on how a program should correctly handle large size
renders, Christoph (and others)?
Thanks,
George
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
This is great, thanks Txemi! :)
George
> Have you tried the zips sent to the IRTC?.
> I suppose you have a lot of scenes there
>
> > The longer rendering time, the better :)
>
> I think the winner image from the Loneliness
> round (made by Gena) could satisfy your needs ;-).
>
> Bye
>
> Txemi Jendrix
> http://www.txemijendrix.com
>
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
