Apple Silicon

So Logic Pro running native Arm with Rosetta 2 plugs won’t work?
I kinda figured it'd be a cluster mess.
 
GaryMcT said:
That’s surprising. If they can translate an app, they should also be able to translate a dll.
Click to expand...
I think the problem is translation would be all or nothing. They can't glue together a native app with a translation of an x86 app. Not to say it's impossible, but I think it's likely to be a mess. From simple stuff like handling the dynamic linking and address space patchups correctly, to things like calling conventions and register management ... not worth the effort for the return I think.
 
GaryMcT said:
That’s surprising. If they can translate an app, they should also be able to translate a dll.
Click to expand...

The juice wouldn’t be worth the squeeze. Like promit says, it’s theoretically possible to translate a component, but by time Apple silicon macs suitable for music production come out, your favorite daw and plugins should be ported to arm anyway.
 
GlennO said:
The juice wouldn’t be worth the squeeze. Like promit says, it’s theoretically possible to translate a component, but by time Apple silicon macs suitable for music production come out, your favorite daw and plugins should be ported to arm anyway.
Click to expand...

You guys know for sure that they don’t translate plugins, or just speculating?
 
Could be that it will actually work, just slowly. I still remember when they switched from PowerPC to Intel. There was Rosetta as well, and nominally it did run PowerPC software, but e.g. Photoshop was unusably slow. Plugins and other high performance software often include rather gnarly low level optimizations, which just won't have the same performance on a different instruction set. For instance, on Intel the maximum width of vector operations is 256 bits (or even 512 bits for Xeon). On ARM it's 128 bits, and there is no way to convert between these instruction sets without a drastic performance penalty.
 
Running under Rosetta will work in some cases, but sorry, no, you can’t mix x86 and arm in a single process, if that’s what you mean by “work”. Compartmentalizing with VEPro should work for VIs, but I have no idea when VEPro will be ported to arm. You can force a ub daw to run under Rosetta though.

Also note Rosetta won’t work with everything. It doesn’t handle my plugins.

More info here: https://developer.apple.com/documentation/apple_silicon/about_the_rosetta_translation_environment
 
Last edited:
I want a Windows machine with one of these so I can run my latest Java programs using an x86 emulator. Twice as slow!!! Or Python!! Yeah, that's the ticket. Let's write everything in an interpreted language and run the interpreter through an emulator. Might as well go back to BASIC.

All these benchmarks mean nothing. Programs 20 years ago were faster than anything nowadays. Everything is bloated, slow and buggy. I have two versions of my schematic capture program. The version from 2000 is fast, smooth and stable. The latest version, which they outsourced to India and ported to Java is a disaster. Takes five seconds to open the property editor on a 3.5 GHz machine. More than five open tabs and it slows to an unusable crawl. Complained to the developers and Javenkasharadipthalon said "You are not using the program correctly, you should not open multiple tabs". "Well why the f*ck do you have tabs if I'm not supposed to use them? And why doesn't the old version have a problem with this?"
 
Last edited:
Don't look now, but I think there are some kids on your lawn :).

Just like last time, Rosetta 2 is a temporary crutch and will be removed after a while.

Music production software was definitely not faster 20 years ago. 20 years ago, desktop computers just barely had enough juice to begin to do realtime audio processing. Audio plugins these days are far more capable because of advances in cpus (and systems) and are actually pretty tight as a rule. It helps that users always have one eye on the cpu meter :).

Is there bloat in other domains? Absolutely, but while music production apps may not be advancing as quickly as in years past, in general they don't have a lot of fat.

I will agree that, for music production, these early geekbench scores might turn out to be misleading. We'll know soon enough.
 
Last edited:
which they outsourced to India and ported to Java
Click to expand...

That's the sad state of the software industry these days. It may look good on paper, but you're getting a lot of these developers straight out of college. I worked for a company who did this thinking it would save them money. In the end we had to toss everything they did and rewrite it in-house with more "expensive" developers.
 
indygeek said:
That's the sad state of the software industry these days. It may look good on paper, but you're getting a lot of these developers straight out of college. I worked for a company who did this thinking it would save them money. In the end we had to toss everything they did and rewrite it in-house with more "expensive" developers.
Click to expand...
History doesn't repeat itself, per se, but it often rhymes....
 
Yeah. Prob not doing the M1 mini after reading you can’t upgrade the ram without paying the ridiculous Apple price and no external GPU support:
“There are some other key limitations with the M1 worth mentioning. Macs with the new chip will not work with external GPUs, Apple says. That’s a particularly tricky constraint for the new Mac mini, which is already hamstrung by its memory ceiling and furthermore can’t be used with a more powerful graphics chip and enclosure combo, as some of have done in the past with Apple laptops and the 2018 and earlier Mac minis. Apple also says the RAM on M1 models is not upgradable — it’s integrated into the system-on-a-chip, so you can’t upgrade an 8GB model with another stick of RAM.”
 
MSS said:
Yeah. Prob not doing the M1 mini after reading you can’t upgrade the ram without paying the ridiculous Apple price and no external GPU support:
“There are some other key limitations with the M1 worth mentioning. Macs with the new chip will not work with external GPUs, Apple says. That’s a particularly tricky constraint for the new Mac mini, which is already hamstrung by its memory ceiling and furthermore can’t be used with a more powerful graphics chip and enclosure combo, as some of have done in the past with Apple laptops and the 2018 and earlier Mac minis. Apple also says the RAM on M1 models is not upgradable — it’s integrated into the system-on-a-chip, so you can’t upgrade an 8GB model with another stick of RAM.”
Click to expand...

I’d bet that these limitations are not likely to go away. If we assume that the current Apple lineup stays the same then except for the Mac Pro I don’t think any system is going to support upgradable RAM, and possible don’t support external GPUs either. The RAM hasn’t been upgradable in laptops in years, and with pushing everything on the one chip they get tremendous speed advantages. The fact that the GPU (which is already faster than fairly recent Nvidia and AMD offerings) is sharing the same RAM I think is a sign of times to come.

Obviously we’re going to get a lot more options in the next round of Apple Silicon chips (M1X possibly) that would be even faster both CPU and GPU, supports a lot more RAM and so on. But I’d say you almost certainly will have to order your computer with what you want, and replace the entire thing if you want more after the fact. The likely exception to this would be the Mac Pro that’s likely the last to transition and will likely have mind blowing performance even in its base configuration.
 
Wow, GPU performance is surprisingly good.
https://www.anandtech.com/show/16252/mac-mini-apple-m1-tested/3
“Overall, the M1’s GPU starts off very strong here. At both Normal and High settings it’s well ahead of any other integrated GPU, and even a discrete Radeon RX 560X. Only once we get to NVIDIA’s GTX 1650 and better does the M1 finally run out of gas”
”Ultimately, these benchmarks are very solid proof that the M1’s integrated GPU is going to live up to Apple’s reputation for high-performing GPUs. The first Apple-built GPU for a Mac is significantly faster than any integrated GPU we’ve been able to get our hands on”
 

Shows a guy running protools on an M1 air In rosetta 2
Seems to run very smooth.
Also, no instruments in the session, which seems to knocks down performance.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom