abstrax wrote:Network rendering: Can be enabled via the application preferences. It allows you to use multiple computers (the master plus one or more slaves) to render one image. Conceptually the slave computers are handled like additional GPUs. Since the slave is basically a special build of Octane Standalone it (currently) requires a standalone license to run. The total number of GPUs (local + remote GPUs) allowed is 12.
Your OTOY NZ Team
The 1 and the 2 are swapped - should be at least 21... PLEASE GUYS!!! Should be easy to fix...
They also said they are going to release a professional level network solution that allows unlimited GPUs. Don't know what the cost for a license is though.
This really makes it better to get one really good card vs a couple cheaper cards. Also, I think it counts the 590 and 690 as 2 GPUs. So you can only have 6 690s on one network. This really makes the 780 look awesome as it has similar performance to the 690 but more ram and only 1 GPU.
I'm really disappointed about the current limit. Perhaps Otoy are trying to somehow protect the cloud offering. When I hear about a 'professional version' I shiver. Small studios that may have 5-10 good machines with potentially 20-30 cards will lose out if the 'professional version' is priced out of reach. They just have not hit the sweet spot I think. It's like giving us a taste and then BAM! - a limit of 12.
Otoy, please make the limit more reasonable. Make the local network offering more accessible, more powerful.
many users like me have 2 or 3 quad gpu workstations.
It's easy to reach a total of 12 gpus
None here own barebones with 8 gpus except big studios.
Never seen big productions with Octane.
In few words there aren't big Studios that are using Octane standalone or Octane for Max or Octane for Maya.
With us Otoy is checking if the distribute rendering works on our poor systems, and we can say "yes".
Now we can render beautiful images with 12 Titans and 3 workstations.
If we want to realize blockbuster movies we need rooms full filled of gpus and great power.
We have to render 2K or 4K frames using about 8 gpus for each frame. Yes, we can buy Tyan barebones.
In this case we need a license for each machine that can handle 12 gpus.
To complete a movie we have to render thousands of frames. We need a minimum of 10 Tyan barebones and a power of 25KW.
Dear Octane user, do you really need 21 gpus to render a teapot?
Gabrielle,
You're forgetting that network rendering gives us potential access to all GPUs on our network and this means we could potentially access all our cuda cores for Octane work, including older cards that might have less cores. Having 12 titans would be great but we don't all have that luxury. So many studios may have a bunch of older cards in machines on the network they can use.
And do you really think your teapot comment is fair? Most users that might be serious about network rendering do a little more than that.
I should reiterate that I'd be happy to pay for more than 12 networked GPU's, as long as it's a reasonable price. And I still maintain that a limit of 12 GPU's is merely one step up from a single standalone license that currently offers support for 7 GPU's at least. Not my idea of utilizing the potential of a network. I just hope Otoy's pricing model to be able to use say 20 local GPU's is nit pitched at the 'movie studio' level. I should not be the only one with that concern.
Finally, Octane offers us speed and quality. But not quite yet full interactivity - interiors are a particular problem. That is the next wave. Brigade style. Computers have been limited as intuitive design tools in my view, at least for architectural work. But Otoy gave us a host of useful plugins. What we need is more speed on large scenes and interiors. Network rendering can give us that. The cloud too of course, but having local network power for everyday, iterative work is crucial.
Will I use 20-30 GPU's on my system if I can. You bet ya! If it's affordable
Seekerfinder wrote:Gabrielle,
You're forgetting that network rendering gives us potential access to all GPUs on our network and this means we could potentially access all our cuda cores for Octane work, including older cards that might have less cores. Having 12 titans would be great but we don't all have that luxury. So many studios may have a bunch of older cards in machines on the network they can use.
And do you really think your teapot comment is fair? Most users that might be serious about network rendering do a little more than that.
I should reiterate that I'd be happy to pay for more than 12 networked GPU's, as long as it's a reasonable price. And I still maintain that a limit of 12 GPU's is merely one step up from a single standalone license that currently offers support for 7 GPU's at least. Not my idea of utilizing the potential of a network. I just hope Otoy's pricing model to be able to use say 20 local GPU's is nit pitched at the 'movie studio' level. I should not be the only one with that concern.
Finally, Octane offers us speed and quality. But not quite yet full interactivity - interiors are a particular problem. That is the next wave. Brigade style. Computers have been limited as intuitive design tools in my view, at least for architectural work. But Otoy gave us a host of useful plugins. What we need is more speed on large scenes and interiors. Network rendering can give us that. The cloud too of course, but having local network power for everyday, iterative work is crucial.
Will I use 20-30 GPU's on my system if I can? You bet ya! If it's affordable
