analogika

Not really. This M2 Pro at least runs pretty cool, and when the fan does come on, it’s inaudible until it hits about 5000 RPM — and then it’s barely noticeable.

Wait — if you’re using the Gain plugin, the gain is applied wherever the Gain plugin is. If it’s the last plugin in the chain, that’s where the gain is. The only difference between a Gain plugin in the last slot and using the fader is that the plugin also affects what gets sent to an Aux bus, even when that is set to pre-fader.

Weird. I haven’t been able to achieve this. Could you post an example of how you did this?

Okay, so why do you think the E-core load drops on the M2 at the very moment Logic playback begins? The only thing Logic appears to be able to split is a few very specific audio instrument plugins — in fact, Alchemy is the ONLY one I can get to work on multiple cores (a maximum of three). The entire channel strip beyond that is always a single core. We have no idea what it looks like on M1 CPUs, because we're only seeing those already under full playback load, unlike the M2 and M3, where we're seeing them at idle AND under load. Interestingly, he doesn't show us this in either of his previous videos on the same subject, either. Both of which are built upon misrepresenting the M2 Pro as somehow inferior to the M1 Pro (which, as repeatedly stated above, is just a lie. He's comparing completely different product tiers). It's quite possible that the M1's efficiency cores looked almost the same just before he hit play, and load on the e-cores did not actually increase from Logic usage at all. Unless you've tried it yourself, I'm not going to accept any claims of M1 performance "seeming to work just fine" from a guy whose main claim to fame is three videos with click-bait titles based on a lie. (Incidentally, those three videos each outnumber the usual plays for his channel by 10x - 100x!)

I'll assume that Logic can throw stuff like meters and display scrolling, notation and the like at the E-cores, since those don't require much CPU. It does not appear to process the actual audio channel strips on the E-cores at all, which is the basic topic of this thread.

I'm not sure that we're seeing what you say we're seeing. On the M2, you can see clearly that the moment heavy processing begins on the performance cores, load DROPS on the efficiency cores. It is extremely unlikely that the efficiency cores are suddenly doing less work — I believe that what's actually happening is that the efficiency cores are being clocked up from their idle speed (around 700 MHz IIRC) to their full 2.4 GHz — to be prepared to handle any overspill from the performance cores, should they get overloaded. AFAIK, this is how Apple Silicon deals with load balancing across CPU cores (for processes whose QoS settings allow assigning them to efficiency cores). Does anyone know whether the CPU History window in Activity viewer shows current load per core as a percentage of power at current clockspeed or as a percentage of power at maximum clockspeed? If the former, that would explain the M2 performance graphs. We cannot compare the M1, since those only show the CPU already under full load. That also doesn't explain why the M3 shows little difference, but that may just be due to the fact that there's so much E-power available in those six cores, that the efficiency tasks just never really register much, anyway. Or they've changed something in the thread handling in M3. I haven't found a detailed analysis yet on my usual source for this stuff, eclecticlight.co (highly recommended, btw.) Because, again, they likely have completely different threading models.

Not in the video you posted at the start of the thread. At 7:55, you can clearly see Logic NOT using the efficiency cores on M1 Pro or M2 Pro. And Logic is supposed to know this how, exactly? Over in the other thread, somebody managed to overload a thread with 18 instruments within a single instance of Omnisphere. How is Logic — or the OS — supposed to assess whether a single thread could overload an efficiency core, and know that it needs to go to a performance core instead? The fact that IN THIS CASE it wouldn't have made a difference does not mean that it wouldn't have to be taken into account. Can you imagine the thread Sascha Franck would post if it turned out that switching on full processing including efficiency cores would suddenly cause certain projects to overload that worked just fine when using fewer cores? I shudder at the thought. 😉 I reassert: threading within Logic would have to be completely re-done, allowing the breaking-up of channel strips into multiple threads (the way Reaper and Cubase apparently do). Why are you posting a video about the M3 Pro and M2 Pro when I explicitly wrote about the M3 AIR and the M2 AIR? You write in your initial post that you're buying a MacBook Air. Why does the core count of the Pro-series CPUs matter to your decision?

It's literally in the video, at second 0:40 — he's comparing a 10-core M1 Pro to a 10-core M2 Pro. Comparing an October 2021 high-end model to a January 2023 entry-level model and being surprised that the earlier machine comes out ahead is just dumb. Had he compared like-for-like models — the 8-performance-core M2 Pro against the 8-P-core M1 Pro he used — the M2 Pro would have come out well ahead, with probably around 95 tracks vs. the 79 of the M1 Pro. All real-world comparisons and all benchmarks I've seen bear this out. I'll note that I'm particularly invested in this exact comparison, because I needed a new machine in January 2023, and this was the exact decision I had to make. There's several threads on this forum where I bounced back and forth all the benchmarks and technical specs with the wonderful folk here. Between the M1 Pro and the M2 Pro as used here? Of course there is! The low-end M2 Pro has 25% fewer performance cores than the high-end version of either generation! But each of those cores is about 20% faster, which directly translates to better performance going from equivalent models of M1 to M2. The only real outlier here is the M3 Pro, which Apple has deliberately limited to fewer performance cores than the previous two models — presumably to get us audio folk upgrading to the Max models, as well (which previously only made sense for video work). "Anymore"? It doesn't look like it did on the M1 Pro or M2 Pro, either. The efficiency cores are seeing quite a bit more load on older systems, but that's largely due to a) their running more slowly (2.1 GHz max vs. 2.4 GHz vs. 2.8 GHz), and b) the fact that there's fewer of them sharing the background load (2 vs. 4 vs. 6). The phrase "so bad" is doing a LOT of work here. 🙂 It's really, really difficult to get an MX to overload doing audio work of any kind… I do hear your complaint that Logic isn't using the efficiency cores. The likely reasoning for this has been discussed plenty in the past; short form: A channel strip is always a single processing thread in Logic, so it's always limited to a single CPU core. Since actual CPU load of a channel strip depends not only on plugins used, but also on polyphony, and additionally, Logic has no way of assessing required CPU load of third-party plugins, it's dangerous to just reassign a thread that works fine on a performance core to an efficiency core, which only has about 60% the maximum performance (as of 2022: https://eclecticlight.co/2022/10/05/making-the-most-of-apple-silicon-power-2-core-capabilities/ ), as it may overload immediately. The solution is to completely rebuild Logic's threading mechanism — a daunting task, I assume, since it also needs to account for all timing issues in real-time processing. I actually believe that allowing Logic to utilise the GPU is a potentially far more rewarding task (however, I am not a DSP engineer). The Good News here may be that Apple hired Universal Audio’s Vice President of Technology and Chief Architect, Dave Tremblay as Senior Manager of Audio Systems. I've also heard rumblings that CoreAudio is getting a revamp in the not-too-distant future… In any case, none of this should affect your decision to buy an M3 MacBook Air over an M2 Air. Any and all DAWs, including Logic, will be faster on the M3 than on the M2.

Since you’re in the market for a MacBook Air, none of this video is in any way relevant to you. It’s only concerned with the number of cores dropping in the M3 PRO (as mentioned, the M1/M2 “drop” is just a lie). That’s not an issue in the Air. The regular M3 has the same number of efficiency cores and performance cores as the M2; they’re just faster across the board.

That part is bullshit. I don’t know whether it’s an intentional lie or just misrepresentation. He’s comparing a HIGH-END M1 Pro with an ENTRY-LEVEL M2 Pro. It only sounds fair because they both have 10 cores. But one of the major developments of the M2 generation was the addition of two extra efficiency cores across the line. He's comparing a six-performance-core M2 with an eight-performance-core M1. A 3000€ M1 machine vs. a 2000€ M2 machine. The entry-level M1 Pro has six performance cores (8 total) and is about 15% slower than the entry-level M2 Pro. Likewise, the non-castrated M2 Pro with eight performance cores (12 total) is 15%-20% faster than the 10-core M1 Pro.

How weird. It wasn't showing earlier, but it's all fine now.

Has anyone got this plugin to work?

I actually wrote that in my post 😉 , but I just checked again: MacOS Extended (Journaled)

I get this a LOT, when I transfer Logic show files from my main studio machine to the show production machines. I used to get it all the time when I just copied the concert folder (about 20 Logic files created in 10.4.8) to my Samsung T5 (formatted MacOS Extended) and onto the show machines (two older MacBooks running Monterey and 10.4.8). That just started happening regularly at some point, but it stopped immediately once I began .zipping the folder before copying it, so I chalked it up to the T5 becoming unreliable. It was never an issue when using AirDrop, but wireless environments are sometimes pretty congested, and AirDrop is really slow on those old machines. Then we replaced one of the machines with an M1 Air running Ventura and Logic 10.7.x, and I now get it every single time on specific projects — only on that machine. The other, still on Monterey and 10.4.8, loads them without complaining. When I dismiss the warning, the projects appear to run fine.

It's back up. FWIW, the regular price on this plugin is 8€. Still downloaded, because I love the DS-1.