Elvissuperstar007 wrote:hi, you will not change your mind to release 2.25?) we project the film is, you need a cloud
Do not take on commercial work based on upcoming versions/features of Octane that are not yet out of beta..
This.
Goldorak, what could be the max. number of gpus running when using orc?
Through amazon.aws it was possible to do 16 (i thought 1 gpu was the equiv. of a 680), is this still possible?
Elvissuperstar007 wrote:hi, you will not change your mind to release 2.25?) we project the film is, you need a cloud
Do not take on commercial work based on upcoming versions/features of Octane that are not yet out of beta..
This.
Goldorak, what could be the max. number of gpus running when using orc?
Through amazon.aws it was possible to do 16 (i thought 1 gpu was the equiv. of a 680), is this still possible?
Best,
ORC can pull in 1000's of GPUs for a single render job. This should be considered for animation and VR output, where this amount of compute power is absolutely indispensable (we have hundreds of GPUs here in LA office, but use ORC to get ~10x more power to do VR renders overnight for some of our partners).
Another ORC case study: light field render job (~1 m cube) requires at least an order of magnitude more time than a GVR stereo cube map of the same scene.
GarageFarm.net offers nice service with 8 x Titan GPUs iirc. Only downside is that you have to install and manage software and licences on remote/rented computer. In other words, while you are rendering on that farm you can`t use local PCs.
v3's improved net rendering will be the cheapest solution for normal static renders under 4 MP. ORC is for jobs that can scale across hundreds to thousands of GPUs and finish months of 10x gpu render work overnight. It will also be critical for punishing massive datasets for real time use in VR/AR and modern post processing pipelines.
Goldorak wrote:v3's improved net rendering will be the cheapest solution for normal static renders under 4 MP. ORC is for jobs that can scale across hundreds to thousands of GPUs and finish months of 10x gpu render work overnight. It will also be critical for punishing massive datasets for real time use in VR/AR and modern post processing pipelines.
You guys have been overpromissing that stuff since GTC 2013. And it's starting to look like this:
It's time to deliver.
Just saying.
Regards
Milan
Colorist / VFX artist / Motion Designer
macOS - Windows 7 - Cinema 4D R19.068 - GTX1070TI - GTX780
Goldorak wrote:v3's improved net rendering will be the cheapest solution for normal static renders under 4 MP. ORC is for jobs that can scale across hundreds to thousands of GPUs and finish months of 10x gpu render work overnight. It will also be critical for punishing massive datasets for real time use in VR/AR and modern post processing pipelines.
Hi Goldorak, is that mean that ORC is not ment to be used by small (freelancer) customers? Only for a big companies and such?
For instance i have Full HD animation, aprox. 300 frames. On my local farm it takes about 7-8 hours to finish rendering, will ORC be helpful to me in any way?
I'll be honest. I have no problem with how long it is taking for V3 and ORC. I understand it takes time to get it all done correctly. I really just want to know exactly how the pricing for ORC will work. We have had some snippets, but as far as I know, nothing concrete.
We are a small company with only 3 dedicated people working on renderings. Our computers are able to render out our normal workload, but every couple of months we get a flythrough animation for an architect and the render time is crazy. ORC might be our solution, but if the cost is high it might be better for us to buy another quad GPU comp for our internal network.
We have been waiting over a year to make our decision because we don't want to buy the new hardware only to start using ORC. ORC would be our preferable method. Makes things easy, dont need to worry about hardware, scales to what we need it, etc. But without comparing cost, we can't make a decision.
It is getting to the point that we might need to buy the hardware just to be able to manage incoming projects which would mean it wouldn't be cost effective to then use ORC.
If you guys wait too long to release ORC, or even just the concrete pricing plan, you might end up losing costumers who need a solution for their growing workload.
