|
Post by kcatthedog on Mar 26, 2021 15:56:26 GMT -6
|
|
|
Post by timcampbell on Mar 27, 2021 15:05:12 GMT -6
Well apparently the m1 version won't run with Protools HDX or native card
|
|
|
Post by kcatthedog on Mar 27, 2021 16:06:30 GMT -6
Would that be apple or avid software issue, as I thought PT was big Sur certified but not m1, how could that be?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 27, 2021 17:33:55 GMT -6
Would that be apple or avid software issue, as I thought PT was big Sur certified but not m1, how could that be? Sounds like Avid may still have a lot of assembly language in their code base. Assembly is a lower-level language that is specific to a particular processor. If they're still using assembly, that part of the code will have to be rewritten for m1. A better idea would be to get rid of the assembly code completely. There's still a feeling in many programming circles that assembly is necessary for maximum speed. While that was true once, I think it's a lot less true now. You may always need absolute top speed out of some functions (let's say an FFT) but that stuff can be encapsulated in a library. In my several-decade career as an engineer, I started out writing a lot of assembly. Probably did a good million lines of it. But in the early 2000s, it was clear to me that both the processors and the compilers were catching up. Generally it's a matter of conceptualizing your task properly and understanding your tools. I never wrote another line of assembly after about 2004. So that's my guess. Avid did a big code cleanup a few years back, but there are probably corners of that code where people fear to tread. Thar be dragons.
|
|
|
Post by svart on Mar 29, 2021 11:11:44 GMT -6
Would that be apple or avid software issue, as I thought PT was big Sur certified but not m1, how could that be? Sounds like Avid may still have a lot of assembly language in their code base. Assembly is a lower-level language that is specific to a particular processor. If they're still using assembly, that part of the code will have to be rewritten for m1. A better idea would be to get rid of the assembly code completely. There's still a feeling in many programming circles that assembly is necessary for maximum speed. While that was true once, I think it's a lot less true now. You may always need absolute top speed out of some functions (let's say an FFT) but that stuff can be encapsulated in a library. In my several-decade career as an engineer, I started out writing a lot of assembly. Probably did a good million lines of it. But in the early 2000s, it was clear to me that both the processors and the compilers were catching up. Generally it's a matter of conceptualizing your task properly and understanding your tools. I never wrote another line of assembly after about 2004. So that's my guess. Avid did a big code cleanup a few years back, but there are probably corners of that code where people fear to tread. Thar be dragons. Not necessarily true these days, as you pointed out yourself. Assembly is still compiled into machine code. Assembly doesn't have to be CPU specific and unless done in an embedded environment, is rarely CPU specific anymore. Last time we did any assembly in anything at my job was around 2008, so a similar timeframe as yourself. We did it because we forced an 8 bit MCU to meet a crazy timing constraint, probably similar to why you needed to use it in your cases too, but there were no core specific commands, just a tight loop with generic instructions. What's probably true is that they did a lot of optimizations for specific CPUs (at a higher level) leveraging things like MMX/SSE/AVX, etc, which don't have direct analogs in ARM.
|
|