OctaneRender® 2023.1 [updated 2023-10-18]

[thorwaldhoffmann]

Great update with a lot of nice stuff.

For the node I have the problem, that even if I select GPU 2, 3, 4 also GPU 1 is rendering on the node.

Is there a solution or is it a "bug"?

Thanks a lot!

[manalokos]

I am getting this error on some scripted operators that worked on previous versions.
Could you fix this?

Thanks

Kind regards
Filipe

[john_otoy]

For the node I have the problem, that even if I select GPU 2, 3, 4 also GPU 1 is rendering on the node.

Thanks for the report. We'll look into it

[john_otoy]

I am getting this error on some scripted operators that worked on previous versions.
Could you fix this?

This error was added intentionally to let users know that calling octane.changemanager.update() from onEvaluate() is not allowed.

We had a scene submitted to us where this was causing random update issues when changing frames.

You can delete that line in the script and it should still work correctly.

[manalokos]

Thanks, already fixed in the script.

Probably unrelated but I am getting a weird behaviour in one of my scenes,
I have a bunch of alembic cameras, and others animated with animate-value lua script, any animation I render with the batch render job
creates duplicate frames, and saves them 2 by 2... very weird.

I'll try to make a example scene if I can

[haze]

Thank you. But I request you add new DLSS 3.5 in next version of 2023.x because new "AI Ray Reconstruction (RR)" will replace AI Denoiser. New RR can fix the bad AI denoiser for caustics at PT on OR 2022.1. :)

But we wait for standalone demo... On today?

Added the demo downloads

[mojave]

Great update with a lot of nice stuff.

For the node I have the problem, that even if I select GPU 2, 3, 4 also GPU 1 is rendering on the node.

Is there a solution or is it a "bug"?

Thanks a lot!

Thank you for the report and apologies for the inconveniences. It looks like some new behaviour was introduced that causes the first GPU to always be enabled in render nodes regardless of your selection.

We have found the cause of the problem and will be fixed in the next release.

[nuno1980]

Added the demo downloads

Thank you! But this version is ~1.08x slower than 2022.1.2 at path tracing kernel with RTX enabled. Why?

[nuno1980]

ATTENTION to my past post.

---------

Benchmark (look "benchmark.orbx" file inside of OctaneBench 2020.1.5) - versions of OR on GeForce RTX 4090 with new 545.84WHQL driver (except not test OctaneBench) running W10 22H2 with CPU Intel i7-12700KF with no E cores:

Code: Select all

	                    Idea | Box specular
2020.1.5 ([email protected])  171  | 169 (!??? where is demo? but OctaneBench)
2020.2		           177  | 131 
2022.1.2	            196  | 149
2023.1		           179  | 135

(in MS/s) at Path Tracing kernel

[gergo_szabo22]

congrats on the 2023.1 guys, analytic lights are a great addition.

one question about the roadmap - is 2023.2 (or 2024.1) with brigade still on the table this year?