OctaneRender™ 2018.1 RC6 Tweet

[Notiusweb]

I now tested Bench 2019 with my 8 Titan X (1 maxwell, 1 Xp, and 6 Pascal), and got 1265.
And...someone got 1298 with 2x 2070 Ti RTX... (in mode "RTX-On")

To me the short and sweet is this....
You will have the same render time as you see now for your scenes, but instead of at the benchmark resolution, you will have that at 1080p.
In other words, if you want to, with a non-RTX card, ball-park simulate what it would be like to have all your cards at RTX speed for a 1080p-ish scene, set resolution at benchmark's 933 x 560.
You can essentially cut the resolution of your scene in half. Whatever speed you get, that is ballpark your card would get on your scene with RTX!!!!

[funk]

I filed a visibility bug in the octane 4 thread but it also affects 2018.




Files are attached here:
viewtopic.php?p=358457#p358457

[pixym]

Hi,
When could I see the "Anime kernel" ?

[PolderAnimation]

I believe when there are more then 5 grayscale vertex attributes the 6th does not work. Is this correct? or is this a bug?

please

[funk]

Ive tried rendering out a few different scenes using cryptomatte (to layered 32 bit EXR).

I've tried the last couple of cryptomatte releases https://github.com/Psyop/Cryptomatte (1.2.3 and 1.2.4), and the reactor version https://www.steakunderwater.com/wesuckl ... =32&t=1775 in fusion 9 https://www.blackmagicdesign.com/au/products/fusion/ .

Every time I select the CryptoInstanceID layer, the script breaks. All the other cryptomatte passes are working fine. V-rays cryptomatte passes also work fine.

There seems to be some issue with the Octane CryptoInstanceID pass.

Maybe this fusion error will give you guys a clue?

Code: Select all

...agic Design\Fusion\Modules\Lua\cryptomatte_utilities.lua:846: manifest string is empty or nil
stack traceback:
   [C]: in function 'assert'
   ...agic Design\Fusion\Modules\Lua\cryptomatte_utilities.lua:846: in function 'load_manifest'
   ...agic Design\Fusion\Modules\Lua\cryptomatte_utilities.lua:731: in function 'initialize'
   ...agic Design\Fusion\Modules\Lua\cryptomatte_utilities.lua:912: in function 'create_cryptomatte_info'
   ...ng\Blackmagic Design\Fusion\Fuses\Matte\cryptomatte.fuse:583: in function <...ng\Blackmagic Design\Fusion\Fuses\Matte\cryptomatte.fuse:568>
Cryptomatte3 failed at time 0


[J.C]

If I use the "veTcron (typo) shadertoy primitive" example and apply a universal material, some faces render black.

Happens in many flat surfaces.
Looks like Universal material bug. In previous release RC4 it was ok.

[divasoft]

Hi, adding portals to the scene significantly increases the noise. Did you plan it this way or is this another flaw?
Cinema 4d integration.

[HHbomb]

As the two renders are almost identical except noise, why do you use portals ?

[Despot]

As the two renders are almost identical except noise, why do you use portals ?

I saw this guys post the other day and thought the very same thing... those renders are almost identical, the area near the central hanging lamp is slightly brighter, that's all to my eyes.

Perhaps he is using them incorrectly.

[haze]

I believe when there are more then 5 grayscale vertex attributes the 6th does not work. Is this correct? or is this a bug?

please

Currently the maximum number of greyscale attributes is 4, and the maximum number of color attributes is 2. We limit this so that it doesn't affect performance too much