Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

ProgRocker

macrumors regular
Original poster
Apr 24, 2018
100
42
I'm currently using a Win10 PC for mostly music production. Though seriously considering switching over to Mac Arm. I have no significant problems with my current setup, however I believe Apple Mx is the future. My question is, what is the significance of the 3nm process ? And what will that bring to my work as opposed to the 5nm process ? Since I am in no hurry, would it be worthwhile to wait for 3nm since it's release seems imminent ?

I'm looking at the Mac Studio Mx Max, and running Apple Logic Pro as my DAW.

Thanks...
 
  • Like
Reactions: oneMadRssn
I'm currently using a Win10 PC for mostly music production. Though seriously considering switching over to Mac Arm. I have no significant problems with my current setup, however I believe Apple Mx is the future. My question is, what is the significance of the 3nm process ? And what will that bring to my work as opposed to the 5nm process ? Since I am in no hurry, would it be worthwhile to wait for 3nm since it's release seems imminent ?

I'm looking at the Mac Studio Mx Max, and running Apple Logic Pro as my DAW.

Thanks...
For music production the current Mac Studio with M1 Max is already more than capable enough. Your biggest gain will be switching to macOS at all. The performance and efficiency gains of 3nm won’t fundamentally change your experience.
 
I'm currently using a Win10 PC for mostly music production. Though seriously considering switching over to Mac Arm. I have no significant problems with my current setup, however I believe Apple Mx is the future. My question is, what is the significance of the 3nm process ? And what will that bring to my work as opposed to the 5nm process ? Since I am in no hurry, would it be worthwhile to wait for 3nm since it's release seems imminent ?

I'm looking at the Mac Studio Mx Max, and running Apple Logic Pro as my DAW.

Thanks...
As @Gudi says, the chip generation is unlikely to make a big difference in the Mac Studio performance. Any performance gains from the current M1x chips using 5nm to a hypothetical M#x using 3nm will be incremental and probably hard to really notice on a day to day basis. If you were getting a laptop, 3nm will likely be more efficient so it might run cooler but that is not a consideration for the Mac Studio which has a massive cooling system.

Besides, we don’t know when Apple will be ready to bring out chips using 3nm and how long it will take them to deploy those chips to each of their models. It make be a couple of years for some models.
 
smaller manufacturing nodes generally provide for chips that consume a lot less power, resulting in potentially better battery life and less heat generated, which as a result will give developers more headroom to create more powerful systems that would have melted before (unless some serious cooling systems were involved) or drained your batteries in no time, unless they are actually aiming specifically for a low power consumption device like a smart watch, in which case the battery life will see noticeable gains.

having said that, the current chips are really fast already, but of course their successors will still be slightly better.
but this will always be the case and the next generation after that will be better than that yet again...
 
  • Like
Reactions: Tagbert
My question is, what is the significance of the 3nm process ? And what will that bring to my work as opposed to the 5nm process ?
Not very much, they're both incredibly advanced manufacturing processes that will have almost no impact on your day-to-day work.

But here is my attempt at an ELI5 explanation.

A chip is made of billions of transistors, where each transistor is a special kind of switch. Like a light switch, there is a point A and a point B such that when the switch is "on", electricity can flow from A to B. When the switch is "off", electricity cannot flow from A to B. Unlike a light switch, changing it from "on" to "off" is different. Instead of a movable contactor, there is special material called semiconductor in between A and B. It works by applying a specific voltage to the semiconductor material between A and B to turn it "on". But this creates heat. And the larger the distance between A and B, the more voltage is needed to turn it "on" and the more heat it makes. Multiply it by billions of transistors, and the amount of heat can be substantial. Conversely, the smaller this distance between A and B is, the less voltage is needed to turn each transistor to "on" and there will be less heat. Thus, smaller transistors use less energy and make less heat.

Very generally, that "nanometer" size of a chip refers to this distance - the size of a transistor. Intel, TSMC, and others all measure it differently and define the bounds of it differently. So the number alone is somewhat vague. But generally, smaller nm process means more power efficiency and less heat to deal with to do the same amount of work.

Last little tidbit: we're just about to reach the physical limits. That semiconductor material is silicon. A single silicon atom is ~0.2nm in size. So right now, assuming 3nm is the size of the transistor, that transistor is only 15 silicon atoms thick. Think about that - counting individual atoms! They've started building "up" now (3D FinFET) to cram more material in a smaller space. But still, we're at the molecular level at this point where pretty soon we just won't be able to go any smaller. Technology will progress, I'm sure there will be other advances to make chips more powerful and more efficient. But it won't be by making transistors smaller and smaller forever because they're already about as small as can be.
 
  • Like
Reactions: Tagbert
Last little tidbit: we're just about to reach the physical limits. That semiconductor material is silicon. A single silicon atom is ~0.2nm in size. So right now, assuming 3nm is the size of the transistor, that transistor is only 15 silicon atoms thick
Nice explanation but this part is mostly wrong. There are no transistors on a 3nm process that are actually 3nm. TSMC calls their process N3 and N3E but the 3nm is marketing.
 
Last edited:
  • Like
Reactions: oneMadRssn
Similar to the same reason why Apple switched to Intel in 2006:

1. Faster
2. Lighter
3. Cooler

Smaller process => more on the chip => more the chip can do.

Back in 2006, PowerPC stagnated and was only focused on faster if that! Intel's offerings and roadmap met all three criteria above, so the switch happened. Now we're looking at custom ARM-ish-like-arino processors for the exact same reasons.
 
Nice explanation but this part is mostly wrong. There are no transistors on a 3nm process that are actually 3nm. TSMC calls their process N3 and N3E but the 3nm is marketing.
Fair point, and yea probably not whole transistors. I've read that the channel length or gate width might actually be 3nm. Either way, we aren't likely to go that much smaller where +/- a few atoms can have major % swings in size.
 
  • Like
Reactions: Tagbert
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.