OctaneRender 2.0 Workstation AMI

[JohnDavis]

Yes, both would be great. But if I'm not mistaken, SDK is the core doc I want.

[JohnDavis]

Mainly the SDK.

[DayDreamer]

Oh boy, where do I sign? I so much want to test cloud Can't really decide if I need to buy Titans or use the Cloud.

[Goldorak]

Oh boy, where do I sign? I so much want to test cloud Can't really decide if I need to buy Titans or use the Cloud.

We'll publish Octane CE+DCC apps on X.IO, which is our app streaming platform. It is going to be way simpler, and ultimately cheaper, than doing this through AMIs in the AWS marketplace.

[DayDreamer]

Thanks for such a quick reply. My concern is: Imagine we have a scene with 200Mb static geometry, 50 animated geometry, 200Mb textures and 600Mb HDRi. It would be nice to send a finetuned, ready for final render, ORBX and go to sleep while all that stuff is being uploaded and instantly rendered.

Now how do I approach scene-developing stage, where I switch different heavy HDRIs around. I imagine, all of them would need to be uploaded and reuploaded sometimes as well, which might end up really a timesinc.

So despite feeling joyful about RealTime-rendering, I have this dilemma about still having to buy a set of Titans to have the ability to assemble complex scenes.

[Goldorak]

Offline and real time cloud rendering are two different things. The latter is simple on X.IO - we just load Octane CE and let you run an ORBX export you've uploaded to one of the cloud storage systems we support.

The former is a task you would want to be able to attach to a specific output node (which is planned for 2.1) and have it do it's thing offline in the cloud. Not too different from how a lua script can generate animations and so on from the drop down menu.

With X.IO we want to let users run DCC apps + Octane entirely in the cloud without needing local GPUs.

Or, you can build your scene locally on your own machine (with whatever GPUs you own) and use the ORBX export -> cloud to leverage the cloud service for final rendering of a specific output node.

Which path you choose is up to you.

[glimpse]

..I have this dilemma about still having to buy a set of Titans to have the ability to assemble complex scenes.

Get a decent GPU (set to be able to work on You rig, without the need of sending information up&down while You're tweaking things up. Just don't spend money for Titans(black), get a 780(6gb) for a fracture of price (600$ instead of ~1k$ for TitanBlack)..& can use the spare to try CLOUD once it will be available.

As far as I know (it was written somewhere here) You'll get 4GB of vram in the cloud. Even if You have a single 6giger in on Your rig You can easily prepare a scene (make a good layout, setup render targets, test out mats, etc..) while powering Your monitor & still keep an eye on usage (so You would not get errors while rendering out in a cloud)..

worth to mention that for the best experience having a dedicated screen card is advisable, maybe even from PRO line (quadro/firePro) if You can take advantage of driver optimisations in Your workflow to jsutify the extra You pay for them.

[DayDreamer]

glimpse » Thank you for that explanation! Your avatar makes me smile every single time! To be frank I'm quite happy with two 780Ti OC and realtime is so addicting, that I'm researching more juicy rigs. This is going offtopic. I wonder if embracing DDR4 RAM and newest CPU along with 3rd and 4th GPUs will serve any performance improvment.

So far 780 with 6Gb havent reached German market.

I imagine it would be wonderful, if Render Team tech writes his thought on newest hardware available from time to time to guide clients away from crazy expensive pieces and towards "his pick".

Even though I've spent some time researching how a computer works, I've no idea if 2x RAM speed (DDRIII to DDRIV) will noticeably affect Octane loading times, which can get quite high.

The plan is to wait for the release of DDR4 capable motherboards, buy the stuff and make a post here about how it went.

[bbeepp]

Offline and real time cloud rendering are two different things. The latter is simple on X.IO - we just load Octane CE and let you run an ORBX export you've uploaded to one of the cloud storage systems we support.
...

Which path you choose is up to you.

Hi, how can I get an Invite Code for X.IO?