Hello World: OctaneRender 4 is here

Tue Apr 10, 2018 11:25 am

I get a failed gpu on ORR4 (shows red) right away after clicking a rendertarget, for example. I`m using a single gtx580 on latest drivers on win 7x64. As well installed the latest cuda toolkit.

Tue Apr 10, 2018 9:00 pm

whersmy wrote:I get a failed gpu on ORR4 (shows red) right away after clicking a rendertarget, for example. I`m using a single gtx580 on latest drivers on win 7x64. As well installed the latest cuda toolkit.

Does the same scene work on 3.08? If not, is it possible to send me the scene so I can take a look? It is entirely possible that we have already fixed this issue for the next release, but I'd like to take a look anyway.

Wed Apr 11, 2018 11:27 am

haze wrote:
whersmy wrote:I get a failed gpu on ORR4 (shows red) right away after clicking a rendertarget, for example. I`m using a single gtx580 on latest drivers on win 7x64. As well installed the latest cuda toolkit.
Does the same scene work on 3.08? If not, is it possible to send me the scene so I can take a look? It is entirely possible that we have already fixed this issue for the next release, but I'd like to take a look anyway.

PM sent

Fri Apr 13, 2018 12:51 am

I have noted a pattern, at current state the denoiser is super snappy and fast at lower resolutions, and conversely is slower the higher you get.
I am finding it's not worth using on a 1080p render, or a 2160p render, because the data collection time + denoiser processing takes longer than a higher sample render.

And yes, you can lower samples, but then conversely the fidelity is lost too. I can get better results with more samples almost universally, so a lot of time it just doesn't comparatively pay off.

So I know it was mentioned work will be done later on getting the denoiser faster.
I would say this feature has such game changing potential, that it should be Priority 1/Top Priority/Priority Numero Uno/King of Priorities/Prioritus Primius .
Anyway you say it....you have a potential monster here, please don't have it sleep for too long...
Release the Kraken!

Fri Apr 13, 2018 8:24 am

Fri Apr 13, 2018 11:00 am

+1 for kraken

Fri Apr 13, 2018 11:25 am

Notiusweb wrote:I have noted a pattern, at current state the denoiser is super snappy and fast at lower resolutions, and conversely is slower the higher you get.
I am finding it's not worth using on a 1080p render, or a 2160p render, because the data collection time + denoiser processing takes longer than a higher sample render.

And yes, you can lower samples, but then conversely the fidelity is lost too. I can get better results with more samples almost universally, so a lot of time it just doesn't comparatively pay off.

So I know it was mentioned work will be done later on getting the denoiser faster.
I would say this feature has such game changing potential, that it should be Priority 1/Top Priority/Priority Numero Uno/King of Priorities/Prioritus Primius .
Anyway you say it....you have a potential monster here, please don't have it sleep for too long...
Release the Kraken!

Personly I think we need more of a good basis before this fancy stuff. I think a proper/better displacement is needed more then things like this. I know it is fun and very exiting but we are missing some basis stuff for a production ready renderer. But that is my opinion.

Mon Apr 16, 2018 7:59 pm

PolderAnimation wrote:
Notiusweb wrote:
Personly I think we need more of a good basis before this fancy stuff. I think a proper/better displacement is needed more then things like this. I know it is fun and very exiting but we are missing some basis stuff for a production ready renderer. But that is my opinion.

right

Mon Apr 16, 2018 8:23 pm

Kraken - Kraken - Kraken - Kraken - Kraken!
Come on Otoy....
You can showcase this. This will boost advertising and be a win for current users...
Win - win.
KRAKIFY THIS MONSTER!

Wed Apr 18, 2018 5:55 pm

Hello Octanight Keepers,

Since the next Octane full version release (V4) is designed to showcase the dominance of our favorite renderer and attract an even greater following, please take into account that the 20 GPU limit ought to be a relic of the V3 past. Even your Benchmark page shows, that at least for Linux users, that 11 GPUs in a single system are a present achievement and, with the V4's added 2 render node union, the GPU cap ought to surely be raised now to, at least, 33 (3x11) GPUs. The GPU count limit ought to be an evolver; not written in stone. It ought to match a reality that is constantly changing. The greatest 3d renderer on the planet deserves to show the whole evolution of its full power.