OctaneRender™ for Blender 2.21.1 - 7.10 Win [OBSOLETE]

[Marten74]

It´s not phantom scatter, but it seems to be a nice little plugin that makes placing instances quite a bit easier. I have not used it my self yet, but from watching the instruction video I think it can be used for a variety of different scenarios. Definitely not limited to the game engine.

Asset scetcher: https://cgcookiemarkets.com/blender/all ... der-addon/

Cheers

::mb

[pegot]

Yes - I've seen that add-on before and have also been meaning to try it.

[mirobimbo]

Asset Sketcher is a nice plugin, have bought too
but it isn't competitor for Phantom Scatter.
Phantom Scatter is much more complex.
And it doesn't solve problem with incompatibility plugin version and Standalone version.

[Marten74]

A question about multiple gpu:s

I just bought a second Titan that I need for reaching a tight deadline.

When I render a scene with 100 samples, one gpu renders in 21.64 seconds and two gpus in 13.85. I thought Octane should scale linear? It is identical cards, but different manufacturers. None of the cards should be over clocked, but maybe the answer is in slightly different architecture or clock speed? Might look deeper in to that later on.

The sad part though is that when I render with motion blur (sub frame) the speed gain is worse.

Same scene with 100 samples, shutter 0.2, motion samples 10. One GPU: 42.45 seconds. Two GPU: 41.40 seconds.

Same scene with 300 samples, shutter 0.2, motion samples 10. One GPU: 1:18.35 seconds. Two GPU: 54.42 seconds.

It seems that when a "render segment" gets to small (10 samples x 10) there is almost no benifit of multiple GPUs. When rendering more samples per segment (30 samples x 10) it seems the benifit starts to get better, but far from linear.

Maybe the answer lies in how stuff is loaded in to memory for each frame? If so, will there be more efficient motion blur calculations for animation in the future?

I´m happy for all suggestions how to get the most out of rendering with two or more GPUs.

::mb

[voon]

Usually, you have some sort of management overhead when you start adding in computing devices. In games, three GPUs tend to deliver the real framerates of calculated two. I would imagine you have a bit of overhead in CUDA as well (same goes for CPU, most applications do not give you four times the speed on a 4-core CPU as a single one). 22 vs 14 seconds doesn't come across as completely wrong to me.

[Marten74]

Hello there Voon. No, actually 22 vs 14 is not catastrophic. Its just that Octane and Otoy have been writing for a long time that Octane render should scale perfectly. Well well...

::mb

[dafassi]

Hey i have some feature request.
While i was working on some scenes using alpha textures and animation-sequences i was missing some features:

1) Transparent textures are not transparent in the viewport. In cycles that works great.
2) Working with animated textures is horrible. There is no animation feedback in the viewport. You will only see that you imported an animation when you hit F12 to render. After rendering with F12 you will see the actual sequence-frame in the viewport. But nothing happens when moving in the timeline.
3) Im missing a way to use different UV-Maps for one material/object to use different projected texture-images.

thanks! have a nice weekend

[Saramary]

Hey dafassi,

About your № 3. You can use the standart Blender way to create several different UV-maps for a object and choose which one of them you're going to use.

[dafassi]

Yes, but you missunderstand me.
I want to use both together!
In cycles there is an "UV Map input node", from what you can choose one of your UVs.
And you can combine many uv maps on one material.
Thats what im looking for to have in octane.

[00Ghz]

I am always amazed by the rate of the constant updates for this OR Blender plugin.