OctaneRender™ 2018.1 RC5 Tweet

[haze]

Hi all

We expected that RC4 would be the last release before stable 2018.1, but we came across a couple more issues and we wanted to implement sheen and coating bump/normal mapping as well in 2018.1. We would like to release these and gather some feedback before we commit to releasing a stable version.

Important: Studio builds now only supports 1 GPU, and no network rendering.

As before: 2018.1 and its XB and RC builds will only be available for customers with a version 4 license while 2018.1 is not yet stable. To try 2018.1 you need to upgrade your version 3 license to a version 4 license. This can be done for free via our web shop: https://render.otoy.com/shop/upgrade.php.

Changes since OctaneRender 2018.1 RC4:

• Fixed "affects alpha" option of specular materials when "keep environment" is enabled (fixed the refraction pass being included in the environment in this case).
• Fixed caustic blur for clearcoat.
• Added sheen and coating bump and normal texture inputs.
• Renamed artistic IOR mode from "Default" to "Artistic".
• Implemented sheen and coating bump/normal mapping.
• Fixed visible on specular for universal material.
• Fixed specular material rendering incorrectly when given 0 roughness and Octane BRDF.
• Fixed node to nodegraph comparison in Lua API raising an error.
• Fixed vertex attributes causing a render failure when used with SSS.
• Fixed float vertex attributes causing render failure with displacement.
• Fixed planes sometimes getting rendered black if there are more than one.
• Fixed crash when opening import preferences for unsupported VDBs.

Downloads for users with a regular license
OctaneRender for Windows (installer)
OctaneRender for Windows (zip)
OctaneRender for Mac OS
OctaneRender for Linux

Downloads for users with a Studio subscription license
OctaneRender Studio for Windows (installer)
OctaneRender Studio for Windows (zip)
OctaneRender Studio for Mac OS
OctaneRender Studio for Linux

Downloads for users with an Enterprise subscription license
OctaneRender Enterprise for Windows (installer)
OctaneRender Enterprise for Windows (zip)
OctaneRender Enterprise for Mac OS
OctaneRender Enterprise for Linux


Happy rendering,
Your OTOY Team

[funk]

Thanks for the fixes + bonus bump/normal for the coating/sheen!

[forward]

Great work on getting 2018 to this state so fast after V4 but I have to say that the sudden change of the amount of usable cards with the studio subscription Version is very bad.

My company bought a one-year subscription version with the given facts on the sales page which still clearly states the following:

"A maximum of 20 GPU’s are available for rendering (4.00 or higher)"

Changing this to only 1 GPU seems very drastic and could be a reason for us to want our money back since we are severly hampered by this decision and would have chosen a different avenue of purchase.
I do not want to come off overly negative but I have to voice our clear displeasure with this as a company.

[PolderAnimation]

I see that the vertex attribute on a SSS and speciular materials work. The thing that is now not 100% accurate is the difference between the normal and the denoised image.
The denoised images looks to have less contrast. Attached the difference. This is almost the same file I already send to Marcus and Alwyn. If this specific file is needed I can provide.
Cheers,

[b.o.p]

Great work on getting 2018 to this state so fast after V4 but I have to say that the sudden change of the number of usable cards with the studio subscription Version is very bad.

My company bought a one-year subscription version with the given facts on the sales page which still clearly states the following:

"A maximum of 20 GPU’s are available for rendering (4.00 or higher)"

Changing this to only 1 GPU seems very drastic and could be a reason for us to want our money back since we are severely hampered by this decision and would have chosen a different avenue of purchase.
I do not want to come off overly negative but I have to voice our clear displeasure with this as a company.

Seconded! seems like the decision that was made was very drastically, our studio just bought another 20 GPUs and now it seems like such a waste...... its hard enough understanding the licencing system here at times, you should be able to use as many GPUs as you like up to around 100 GPUs, similar render engines work this way, but holding us to ransom over how many GPUs we can utilise may put more people/studios off taking the plunge with octane its such a shame as it is a great engine. if it stays this way we will migrate back to our previous rendering solution. I would guess the official line will be (4.00 or higher ) meaning only version 4 - version 4.09 ect and 2018 is a different version. so our last set of terms doesn't apply. sorry guys but this news is such a bummer.

[Goldorak]

It’s just temporarily at 1 GPU, the stable release is going back to the normal 2 GPU limit after we finalize changes to the stable and enterprise tiers. We’ve only recently experimented with more than 2 GPU’s for V4 subs for a few months, and while that won’t go away for V4, it’s not going to be applied to later versions as the FAQ states, until we consider what affect NVLink and especially RTX have on the delta between studio and enterprise tiers.

[mr_d_wilks]

It’s just temporarily at 1 GPU, the stable release is going back to the normal 2 GPU limit after we finalize changes to the stable and enterprise tiers. We’ve only recently experimented with more than 2 GPU’s for V4 subs for a few months, and while that won’t go away for V4, it’s not going to be applied to later versions as the FAQ states, until we consider what affect NVLink and especially RTX have on the delta between studio and enterprise tiers.

Please keep in mind that if we don't have RTX or nvlink we would still be limited quite drastically on studio licenses compared to v4. And the wording on the page when we bought in to studio was up to 20 gpus. there was no word of 2 card limit at that time - please keep that in mind. I agree that RTX and nvlink are an unknown right now and could potentially hurt your sales if its a complete game changer. but if we are on cards without these technologies I think it would be very reasonable to have more than 2 cards. For example i have a 4x 980 ti machine that i built after the black friday sale to use with studio license and now when i move to 2018.1 my overall performance is much slower since the OB score from v4 to v2018.1 is only a few percent and i'm limited to 2 cards in the end.

I hope you all will agree to make pascal, maxwell, and kepler cards have a higher limit. 4x GPUs is reasonable for these older cards?

Thanks,
doug

[Goldorak]

It’s just temporarily at 1 GPU, the stable release is going back to the normal 2 GPU limit after we finalize changes to the stable and enterprise tiers. We’ve only recently experimented with more than 2 GPU’s for V4 subs for a few months, and while that won’t go away for V4, it’s not going to be applied to later versions as the FAQ states, until we consider what affect NVLink and especially RTX have on the delta between studio and enterprise tiers.

Please keep in mind that if we don't have RTX or nvlink we would still be limited quite drastically on studio licenses compared to v4. And the wording on the page when we bought in to studio was up to 20 gpus. there was no word of 2 card limit at that time - please keep that in mind. I agree that RTX and nvlink are an unknown right now and could potentially hurt your sales if its a complete game changer. but if we are on cards without these technologies I think it would be very reasonable to have more than 2 cards. For example i have a 4x 980 ti machine that i built after the black friday sale to use with studio license and now when i move to 2018.1 my overall performance is much slower since the OB score from v4 to v2018.1 is only a few percent and i'm limited to 2 cards in the end.

I hope you all will agree to make pascal, maxwell, and kepler cards have a higher limit. 4x GPUs is reasonable for these older cards?

Thanks,
doug

We’re open to any feedback - nothing is locked in yet, and your point is valid.

[coilbook]

will this version or 2019 have a procedural displacement that doesn't need uv mapping and baking?

[Goldorak]

will this version or 2019 have a procedural displacement that doesn't need uv mapping and baking?

2019.1 (next version) will introduce a new vertex displacement system, alongside the older voxel version - the new system uses more memory (it auto-subdivides the meshes it is applied to), but otherwise has no other limitations and is much faster