[Beowulf] 3d rendering cluster

[Robert G. Brown]

On Wed, 25 May 2005, Paul K Egell-Johnsen wrote:

> And with the sound cards one need to have audiodistribution to the
> DAW. A cheap Windows XP lisence cobbled with something called FX
> Teleport will enable 32bit 192KHz 8 channel audio from each audio
> renderer to the DAW over ethernet. The win lisence and that software
> is much cheaper than, and more practical, getting the sound
> distributed by dedicated audio wires.
...
> There are lots of good Linux audio soft, but I'm not going edit any
> samples or sound or try to parallellize that. Instead I'm trying to
> get enough simultaneous playback channels in order to create a
> convincing symphonic orchestra for scoring the animations we're doing.
> 
> One hope, though, is that those libraries are shipped with something
> called Kompakt, a  player. It is both standalone and VST plug ins, and
> I learned of a specialized distro which will allow VST plug ins  to be
> used under Linux. Now that sounds interesting, if I can get the audio
> streams sent to the DAW without adding extra wires.
> 
> I think I'll loose some functionality, the ability to remote control
> the playback from the DAW, but I'd be saving at least 50% on OS cost
> for this (half the Windows price) + the full price of the windows
> soft.
...
> I'm not convinced that we can drop windows on the workstations yet,
> except where we can run Mac OS X of course, the other graphics artists
> are very much invested in 3ds max, which is only Windows - fortunately
> the mental ray renderer is properly multi platform, unless I can find
> a convincing open source alternative. I'd think those often are just
> as advanced as the commercial ones, most of what is new in rendering
> are presented at siggraph and simular open forums and the papers are
> often available as well.
> 
> Thank you for your comments Robert.

De nada.  Sounds like you are very much on track.  It's really
interesting to realize that multichannel audio rendering is yet another
perfectly legitimate parallel application for a cluster (one I had not
thought of in those terms).

Does the orchestral mixing have to be delivered in "real time"?  Can't
you just render the channels separately and asynchronously on the nodes
and mix them on a head node?  Or is the creative process heavily
interactive, so that you constantly need for all the channels to be
rebuilt/remixed?

If the former, then you have a lot of ways to do the rendering and
delivery on top of linux (using much the same sort of process you'll use
for rendering/delivering/assembling frames).  If the latter, hmmm, it
doesn't sound like it would be THAT difficult to write a parallel app on
top of PVM or MPI that would farm out the scores to nodes for rendering
and receive back the results in chunks for mixing and delivery.  I'd
also expect each node to be fast enough (most likely) to deliver
multiple channels on its own -- you have billions of cycles per second
for computation compared to sampling rates in the hundreds of kilocycles
per second, or order of a few thousand instructions per sample to
dispose of.  Just how much computation IS required per channel?

Regarding the 3ds max and workstation issue -- there I understand.  A
mission critical application is just that.  I have no idea if linux has
anything competitive in this arena.  OUTSIDE the critical app pathway,
though, linux on the desktop is currently pretty darn complete and
functional, although one can still get into a bit of trouble with Office
compatibility with really complex documents (Open Office handles simple
ones pretty perfectly, to my experience, but can barf on ppt slides
containing e.g. embedded spreadsheet graphs actively connected to a db
or spreadsheet or word processed document, and sometimes gets things
like graph scales wrong).

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu