Does Arnold work with new M1 Max Chip?

[mlove15]

I just got a new MacBook Pro with the new M1 Max chip, for some reason I can't get Arnold license to work. Is Arnold supported on the M1 Max chip? Because it's not working for me.

[Stephen.Blair]

You have to run Arnold under Rosetta 2

[mlove15]

@Stephen Blair Not sure how to run Arnold in Rosetta 2. No real answers on your support site.

[mlove15]

After using Arnold on Rosetta for a while, it's not worth it. Arnold on Rosetta takes up so much memory. It's sad that they continue to be so far behind, will switch renderers after my current project.

[thiago.ize]

I'm surprised running under rosetta would use extra memory. The data being stored in memory should be the same whether the executable is running native or under rosetta. I would suggest asking a new Arnold Answers question about this so that someone can help you identify why so much memory is being used. If you do that, include a complete "info" level log file so we can see where memory is being used.

[mlove15]

@Thiago Ize isnt the point of upgrading to an arm chip to have better processing and power usage, what’s the point of using an intel renderer on an arm Mac? Especially, when Arnold has been slower than other renderers to update technology.

For me I have to open a blank window and allow the IPR to go before opening my project file, otherwise it won’t pickup the license for some reason.

Sometimes when I use the IPR i get a random spike in memory consumption, sometimes randomly going over 100 GBs.

Plus random crashes and freezing. My scenes can get pretty complexed, but that’s why I would want a renderer that takes full advantage of what Arm chips have to offer.

At this point it just seems more logical to learn a new renderer.

[thiago.ize]

Maybe the memory leak is related to the following possible macos bug memory-leak-bug-is-killing-macos-monterey-performance ? It could very well be something else, but without verbose arnold log files I'm afraid there's not much else I can suggest about what is going on.

The power savings you're already getting regardless of whether Arnold is running natively or not. The performance hit of rosetta is about 21% according to m1-chip-emulating-x86-benchmark/ (so a 10s render under rosetta is 8s if it was native).

[thiago.ize]

The crashes and freezes could be due to the version of arnold you're using. Arnold core 7.0.0.0 has some serious bugs that can do this. 7.0.0.1 or the previous 6.2.1.1 should be more stable.

For the licensing, that does sound like buggy behavior. I would suggest making a new arnold answers question about this and someone familiar with licensing might be able to help you.

We are actively working with Apple on a native apple silicon port, but getting this working without bugs will take time due to all the low level optimizations in arnold, so I can't say when or if this will be available.

[cathy_easthope]

I want to use the power of the M1Max chip with Maya. Can anyone suggest an alternative renderer to Arnold which will run natively on the M1Max?

[thiago.ize]

Hopefully I'm wrong, but my understanding is that even if the renderer runs natively, it can't be used until the host application (Maya in your case) is also running natively.

[Anonymous]

Redshift is native to Apple silicon

[rico]

I switched to Redshift. The Arnold plugin is more feature complete and the workflow is more streamlined. But I think a new Arnold version for Apple Silicon and Metal is not in the pipeline.

[Stephen.Blair]

Arnold for Apple Silicon is in the pipeline.

[Anonymous]

@Stephen Blair I know you multiple times said there is no ETA on this. But it would be cool if there is an indication when we can expect Apple Silicon support Q1/Q2/Q3/Q4?. The waiting is killing 🙂

[jordan7SQ7S]

Been waiting for Apple Silicon Support also. 2 years into this transition and nothing. Meanwhile Octane, Redshift - hell even Chaos Corona announced that M1 is fully supported in Corona 8 with a 50% increase in speed today over non M1 optimized rendering. 

My primary work is done on a PC with a 3990x 64 core cpu and 8 RTX 3090s spread out over 2 nodes so I’m not too salty over the lack of Apple Silicon support. That said I have found doing look dev when I’m on the road or on set is pretty handy using Redshift on my 16” M1 Max. Would prefer Arnold but I’m losing faith that this is going to happen anytime soon. 

[infoHVXV8]

I too desperately need to know when the M1 support is coming. Cinema 4D R26 won't run on Rosetta so I'm stuck with R25 and as crazy as it may sound, I really do need to go to 26 for technical reasons. So when can we expect it???? 

Thx

[leandro.cattani]

And the silence from Autodesk is going on......

[Stephen.Blair]

As I've wrote before, we cannot give out any specific dates or ETA for new features.

We're working on adding M1 support, and it's not Arnold that needs to support M1 chips, we have other dependencies and they all have to support M1 too.

[JohnDraisey]

The Dependencies needing to support M1 first makes sense. I'm interested in M1 now that I've personally seen Noice perform 40% faster on an M1 Mac Mini versus a Ryzen 3950x (saved me this week). Rosetta 2 mode feels good enough for now. But Apple also showed some design limitations with their Max vs Ultra M1 chips. Could be a year before proper M2 Pro and Max chips arrive. If Arnold becomes native by then, thumbs up.

[monotraum]

i am testing and learning Arnold in R25 under Rosetta. Its an ok setup to get to know Arnold (a great renderer!) but the main Rosetta restriction is that Arnold seems to not make use of the 32 M1 GPU cores. Only the 10 CPU cores are sweating. My naive calculation tells me that render time would go down to a 3rd or less. So basically we all are waiting for the day the stallion can run free and have its great day 😉