I’m using SpectraLayers Pro 12 on Windows 11, and I’ve noticed that even when the AI Processing Device is set to GPU, the majority of the actual processing still seems to be handled by the CPU.
I monitored the system via Task Manager and observed high CPU usage and very low GPU activity, even during intensive AI-based operations.
System specs:
CPU: Intel Ultra9 285K
RAM: 64GB DDR5
GPU: NVIDIA RTX 5060 Ti
OS: Windows 11
While I understand that some minor inconsistencies and bugs may be addressed in future updates, the current lack of effective GPU utilization is impacting workflow in noticeable ways. I would greatly appreciate if more comprehensive GPU acceleration could be implemented as soon as possible.
Additionally, I’m using SpectraLayers as a Nuendo Extension.
I upgraded from SL11 to SL12 via installer, and now only SL12 appears as active within Nuendo.
(Both SL11 and SL12 still launch correctly as standalone applications.)
Is it correct to assume that SpectraLayers Extensions in Nuendo do not support backward compatibility (e.g., two versions coexisting)?
I’m a strong supporter of SpectraLayers and very optimistic about its future.
Thank you for your ongoing development and improvements!
What about Apple’s M Series chips?
Considering their architecture, it seems SpectraLayers could potentially take advantage of both the high-performance CPU cores and the powerful integrated GPU. Do you have any insights on whether the software is optimized for that?
Apparently GPU acceleration is not supported for Mac computers. There is no setting to allow it. I saw somewhere that Avid claims that it’s Apple’s fault because they do not provide the information needed to allow Avid to implement it.
Similar claims have been made here (Apple not sharing information) which, as I’ve also stated here, doesn’t make much sense to me. Maybe it’s a generalization that gets the “not yet” point across without going into too much detail about development and economic requirements, impact, and roadmapping. But:
While I would love to see native Apple Silicon GPU support, I understand the market may not be there yet. Even today, many apps are CUDA-only, or take a dependency on general subsystem libraries. Nvidia was on par yesterday to become the most valuable company in human history, so it seems like a solid development bet.
I tend to capitulate to the “we’ll get it when doing so either generates a net profit for the company, or not having it loses customers to competitors that do” mindset.
If I HAVE to have GPU acceleration, then I have the choice of setting up a new system. For me, I can just schedule workflow around it - but I’m not making a living using SL like many of you guys are.
Thank you for such a well-thought-out comment — it really helped me understand the bigger picture. I’m just a tech enthusiast, not someone who works professionally with audio, but I have a lot of admiration for tools like SpectraLayers and the team behind it. The level of precision and the kind of workflow SL enables is truly impressive.
Regarding GPU support on Apple Silicon, I had assumed that, given the power and integration of Apple’s architecture — especially with Metal — this kind of support might already be more common in tools like SL. But I completely understand the points you raised, especially about economic viability and the current dominance of CUDA/Nvidia in the pro market.
I really hope that as the market continues to evolve, we’ll see more native support for Apple Silicon. I trust that, when the time is right, the SpectraLayers team will know exactly how to implement it in the best possible way.
In the meantime, I’ll keep learning, experimenting, and enjoying everything this amazing tool already offers.
Indeed, and the question is really “when,” I hope. To me, this is where the Forced Adaption model (from Apple, not SB) is at its worst, because really, you and I should have the option of integrating something like a Thunderbolt eGPU box to better integrate with other currently-supported GPU solutions, but MacOS actually prevents this entirely, which I think is a bit lame. I mean, I’ve already purchased my M3 Ultra, so it’s not like they’re LOSING money, and since this would be an Apple-Silicon-only-product it shouldn’t really interfere with the adoption model - rather, it gives us options NOW. That said, “options now” sometimes have a way of getting in the way of forcing a future path so…
Totally agree — having the option of using an eGPU with Apple Silicon would make a big difference, especially for those of us who want more flexibility without abandoning the ecosystem. I get Apple’s long-term vision, but it’s frustrating when it limits what we can do now, especially with such capable hardware already in our hands. Hopefully, things will open up down the line. Thanks for the insight — really appreciated!
