I have the following computers available for use.  Does anyone have ideas
about how I could go about doing distributed rendering on them?

400Mhz Pentium II
166Mhz Pentium
33Mhz 486Sx
Mac 68040 (33 Mhz, I think)
Commodore 64
The disk drive from the Commodore
TRS 80 model III
Nintendo

Thanks,
Mark

Post a reply to this message

Hmmm. I think you can get the first two to run PVM POV-Ray. The others...
Well...
(does the 486 have a co-processor?)

Or, you might look into the R-Cast system I'm working on now (it's in
limited beta on the prototype). By the time it's done, that might be able to
utilize the Pentiums and the mac.

Mark Wagner wrote:

> I have the following computers available for use.  Does anyone have ideas
> about how I could go about doing distributed rendering on them?
>
> 400Mhz Pentium II
> 166Mhz Pentium
> 33Mhz 486Sx
> Mac 68040 (33 Mhz, I think)
> Commodore 64
> The disk drive from the Commodore
> TRS 80 model III
> Nintendo
>
> Thanks,
> Mark

Post a reply to this message

Mark Wagner <mar### [at] gtenet> wrote:
: I have the following computers available for use.  Does anyone have ideas
: about how I could go about doing distributed rendering on them?

: Commodore 64
: The disk drive from the Commodore
: TRS 80 model III
: Nintendo

  I don't think that povray is available for those.

  I once used a computer lab for distributed rendering with povray. The
method I used was this (well, actually it was not exactly this at the
beginning, but I soon discovered the right method):
  Divide the image into groups of, for example, 50 lines. Start rendering
a group of lines in each computer available. For example, if you have 4
computers available, you would calculate lines 1-50 with the first one,
51-100 with the second, 101-150 with the third and 151-200 with the fourth.
  Every time a calculation is done, name the resulting file with the
appropriate number (for example file1.tga for the lines 1-50, etc) and
start calculating the next group of 50 lines (eg. 201-250).
  After the last lines have been calculated, you have a bunch of files.
Now you only have to join those images. You can use a paint program or
make a specific program for that (uncompressed targa is a very easy format
to read).

  Note: This method doesn't work if you use radiosity.

-- 
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):5;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/

Post a reply to this message

Nieminen Mika wrote:
> Mark Wagner <mar### [at] gtenet> wrote:
> : I have the following computers available for use.  Does anyone have ideas
> : about how I could go about doing distributed rendering on them?
> 
> : Commodore 64
> : The disk drive from the Commodore
> : TRS 80 model III
> : Nintendo
> 
>   I don't think that povray is available for those.
> 
>   I once used a computer lab for distributed rendering with povray. The
> method I used was this (well, actually it was not exactly this at the
> beginning, but I soon discovered the right method):
>   Divide the image into groups of, for example, 50 lines. Start rendering
> a group of lines in each computer available. For example, if you have 4
> computers available, you would calculate lines 1-50 with the first one,
> 51-100 with the second, 101-150 with the third and 151-200 with the fourth.
>   Every time a calculation is done, name the resulting file with the
> appropriate number (for example file1.tga for the lines 1-50, etc) and
> start calculating the next group of 50 lines (eg. 201-250).
>   After the last lines have been calculated, you have a bunch of files.
> Now you only have to join those images. You can use a paint program or
> make a specific program for that (uncompressed targa is a very easy format
> to read).

  For another angle, at work we have 500 odd desktops sitting idle most 
of the time (disaster recovery). I've managed to setup an automatic 
linux network install that creates mosix nodes. Now I'm working on
scripts to chop up an arbitraty povray render into n-sized chunks to
render. With mosix you just create n instances of povray and it handles
moving the processing around to idle nodes.

   Of course I'm more interested in creating huge animations using this
method. Considering if each frame takes 1min to render (nice easy 
animation) then with 300 nodes a 15minute animation (27000 frames)
will only take around 90 mintues give or take network overhead.

   So far I'm only up to 60 odd nodes but even that renders most of
the povray3.5 test images less than 10mins (barring the real long ones).

-- 
rya### [at] carelessnet

Post a reply to this message