Hi Moogy, I'm not Graphics savvy, but I think the idea is interesting. In the P2P network, artists who have licensed graphics softwares can legally extend their tools to others who don't have?

I understand what you try to do but at least you should specified the plugins by describe it a little bit more. Fresh idea and hopefully this is a new one.

Have a look at BURP ( http://burp.boinc.dk/ ) . It is based on BOINC, a vast distributed computation system famous for its SETI@Home projetc. It currently supports the Blender renderer only, but designed to handle others.

I like your idea of a computation network what would allow people to get paid for their processing. IMHO, it should be more general than rendering. Why not a kind of BOINC+PayPal system that would rely on existing stuff ?

Moogy.
This is a very interesting idea. I'm not sure how easy it is to pull off and whether you can generate enough money. It needs some refining.
Tommy

Very nice idea; I'll vote on it when I can...

Voting starts in: 2 hours, 47 minutes

Thanks for all the comments

I know of BURP. But it doesn't support most of the application the industry uses.

But the main idea is same... just the business model is different.

]V[oogy

It sounds like it could possibly work.. I think the hard part will be making it such that the client doesn't know what it's working on, but it might be possible. I think you would also need to consider that some rendering jobs need a LOT of data to be sent (textures, models etc) to complete it.. maybe you could leverage this somehow to send blocks of frames to be rendered?

It sounds like a good idea, but I agree that it should be more generic, or at least become more generic as it develops. Are there free rendering engines available? If not, the market for 3D rendering is toooo small!

One problem is the economic balance. Are people willing to pay more than the cost of power / bandwidth for processing power? I don't know this industry very well so I wouldn't have a good guess. If people are willing to pay more, why not do it?

With all of these computers running and consuming power, we could offer a carbon-neutral feature for extra!

Well I my initial concept (I'm not a professional programmer). I wanted to brake down the jobs it to groups of pixels 1, 4 ,9 no more of the total image.

The rendering user would start his process on his local system (parcing) (loading: texture, geometry, animation,etc.). The Parcing process can only be done on a single cpu or single tread (that I know of). Once this is done the plug in would give the illusion of having hundreds of CPU's. Each section of pixel would be encapsulated then sent out to the available CPUs.

Hope this makes things more clear.

]V[oogy

if you've done any rendering, you will know that all the nodes will require access to some regio, or possibly all of the dataset. How do you propose to do that? People will likely not want to give away there dataset! How is this better or cheaper than a duo or quad core cpu running at 3 GHz+ with full access to 800 MHz sitting in front of you?

That the idea behind the encapsulated data. You are not accessing the dataset directly.
How about 1000 2GHz cpus... thats better then a quad core.

PS if anyone has a quad or dual core... they know that during the parcing, only one part of the cpu is used...

]V[oogy

Oh good sir

I've been waiting for something like this.
I render alot of things on a daily basis and I think (if i understand this correctly) that sharing render time would be great.

Good idea - it could benefit from my "Project Aldeburgh" coming soon to CH. Watch this space.

yeah, 1000 2Ghz CPUs with only 128kbit upstream between them - that's 16 kb/sec. - wow that's really fast! That's MUCH faster than a few cheap 300$ machines with gigabit ether
(many 100$ motherboards have this now).
As for parsing time, there are parallel implementations, and parse time is often quite small compared to render time. BTW it's PARSING, not PARCING.

Nice one.... you make me think more of other IDEA possible!