Can confirm.

@Clément Foucault (fclem) with the new code it will keep the depsgraph from frame to frame for Eevee, but this seems to be failing with motion blur.

One workaround we have been using is to use emitters instead of hair, perhaps this won't work for many uses of hair but for us it seems enough.

I have the same problem with a Radeon (TM) R7 M340. using Adrenaline 2020.

Can we expect light linking for Blender 3.0? it's really a important feature.
or this GSOC proposal is the only hope? https://devtalk.blender.org/t/gsoc-2021-proposal-idea-light-linker-light-mixer/18185/14

I suppose this is the same problem reported in T85078: OptiX GPU rendering artifacts with large plane

Thanks for the report.
I can confirm it now. (The steps were not clear before).

Yep, solved--ty @Brecht Van Lommel (brecht) !

Double usage of the name property is indeed weird/problematic.
So whenever you change the node name, you are actually changing the name of the refrered AOV.

Caused by rB50782df42586: Render: faster animation and re-rendering with Persistent Data

Please test on a recent Blender version.

Tried to redo on b9207fb43d6fb55b83ee25263a3963c8c736f876, macOS 10.14, Intel HD 6000. No freezes.

@Aaron Carlisle (Blendify) sorry for ping again, requested changes done

Thanks for the report. I can confirm.

Can confirm.

I can confirm.
If you bake the normal tangent in edit mode the result is correct.

2.82a seems to was fine, now I´m using 2.90.0 which seems to work ok too (what happened ??)

FYI, filed this ticket for OSL: Crash in optimize_all_groups running in pre-release Blender on Apple Silicon. Not 100% sure its OSL, but hoping for guidance.

I just wanted to clarify that this is not limited to vertex colors. Thanks to a recently added feature, custom attributes created in geometry nodes can be used directly in Cycles without the need for vertex colors: https://developer.blender.org/rB3a6d6299d
Maybe the fix will apply to GN as well, but I think it would be good if you could modify the name or description of this task to indicate that it also affects geometry nodes, since you closed my task.

Apparently this is related to built-in attributes as seen in https://blender.stackexchange.com/questions/14262/what-can-you-call-from-the-attribute-node#14267

I can confirm the crash on M1 with the Arm 64 official build.

It is still not working (latest build from 14 apr)... really wish to switch for 2.93 because of Geometry Nodes, but this problem completely break workflow :(

@gabe2252 (gabe2252) , thanks for testing!

I still have this problem, if that helps.
(Sapphire Pulse) RX 5700 XT; Blender 2.92; Default scene, viewport just won't work. Although, rendering will work.

Changes requsted except subpanel as subpanel depth is a problem, also they can't be reordered if they become subpanels.

In D10505#279580, @Aaron Carlisle (Blendify) wrote:

Was there a reason you moved the draw_modifier functions to there own function?

Was there a reason you moved the draw_modifier functions to there own function?

This is getting a lot closer from my POV.

Alright, I tested the @thomas-29 (thomas-29) 's blend file and can confirm that the new volume objects do not work.

Oh, I was thinking objects with cycles volume rendering. I haven't used the new volume objects and I didn't realize that's you where talking about that. I'll try your blend file in few hours when I get back to my computer. Feeling a bit embarrassed.

I can confirm this, unless we're both doing something wrong, it looks like this doesn't work at all.

@gabe2252 (gabe2252)
could you please double check rendering using Cycles and OpenCL with GPU selected in the preferences as well as in the render setting with this file:

In T86610#1145496, @thomas-29 (thomas-29) wrote:

The crash is gone, but for the cost of not rendering volume objects on the GPU with AMD OpenCL.
(...)
Would you mind re-opening?

The description in the report does not mention volumes, so this appears to be a different problem.
And I'm not sure if it's really related to the fix.
CC @Brecht Van Lommel (brecht)

The volume objects are rendering for me.

I tested blender-2.93.0-09d7d88cc42a-linux64 from https://builder.blender.org/download/

The issue still occurs. Tested with Blender 2.93 hash 09d7d88cc42a and Blender 2.92. Here is a screen capture in case it's helpfull. I have 2 GPUs and for the test purpose I disabled one of them in the preferences.

This is a limitation of the algorithm indeed.

Alright I tested version: 2.93.0 Alpha, branch: master, commit date: 2021-04-11 20:10, hash: 09d7d88cc42a with my RX580 and it's seem to be fixed.

The problem is still present for me in both 2.92 and 2.93 Alpha 09d7d88cc42a

I was able to replicate the problem the first time today with the Default Cube.

Split kernel error: failed to load kernel_holdout_emission_blurring_pathtermination_ao

But after the kernels compiled again when reloading the file, the problem disappeared.

Can anyone confirm that they still have this issue? I got no clear replies so far. I can't reproduce any issue with my RX 5700 XT card with latest master.

rB91c44fe88547: Cycles: disable NanoVDB for AMD OpenCL may have solved the problem, you can try the latest daily build.
(If you can't, no problem, we can take it as resolved)

I believe that this bug has already been fixed by rB91c44fe88547: Cycles: disable NanoVDB for AMD OpenCL.

@Philipp Oeser (lichtwerk) yes this is currently not supported and should be hidden.

I cannot reproduce this anymore, looks like it is fixed.

Don't know if this is the same issue, but I am having a similar problem, a large scale scene with a character far away from world origin

Those actions (specifically disabling "Adaptive Sampling") cured the destroyed cycles renders in 2.9.2 (Mac OS Mojave).

Sorry, but latest build STILL gives compilation problems. I assume that build already contain commit?

I also had spent a few hours trying to figure out why cycles were hanging when I tried to use both my 1080ti GPUs. Just before I found this thread I did stumble into turning off my hardware scheduling and things started working. I was worried one of my GPUs had failed at first but I tested them in another 3D app and they seemed to be fine.

This feature is sorely lacking

Hopefully a631a9eb1f0b solves this for you @Ilja Razinkov (IPv6).

For me it's in all frames, and with any other project too, but this one was one of the simplest I had with non-instant rendertimes.

I can confirm in some frames:

Can confirm.

Can confirm, also present in 2.93 rB074b0b6da0a4.

In T71258#1143480, @Armand (tsukirockstar) wrote:

@Tamás Péter Radványi (RTamaasP) This is not a place to talk about third party software. The OP (and so Am I) are on Linux. I'm aware about the eternal warning that substance painter gives you when you start it but it not applies to this issue. Furthermore, ATM I'm stuck with no GPU at all at 2.92. You can check that here. As I said, third party addons will not help to solve a bug.

T86780

@Tamás Péter Radványi (RTamaasP) This is not a place to talk about third party software. The OP (and so Am I) are on Linux. I'm aware about the eternal warning that substance painter gives you when you start it but it not applies to this issue. Furthermore, ATM I'm stuck with no GPU at all at 2.92. You can check that here. As I said, third party addons will not help to solve a bug.

So, for what it's worth, I changed this line: https://git.blender.org/gitweb/gitweb.cgi/blender.git/blob/HEAD:/intern/cycles/render/osl.cpp#l171

Using Radeon Software 20.10.1, with MSI rx580 8 gb OC ed. and doesn't work with ANY versions of blender, EXCEPT 2.92, but now it works extremely well. Like before this took literally minutes even to react anything after I applied the volume (and then crashed at 1 samples), now not even increasing the viewport render time.

Unfortunately, this did not work for me either. I also tried Blender 2.93, the same error occured. I use 2 Radeon RX 580 GPUs and the dirver version is 21.3.1. The same thing is with 21.Q1.1 Pro drivers.