Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Chipmaker Discusses Highly Advanced 1.4nm Chips for First Time
- Thread starter MacRumors
- Start date
- Sort by reaction score
Still all these baby steps. Why do they bother with all these in between numbers? Why not just go straight to, say, 0.9nm or 0.5nm instead of mucking around with 1.2nm? They know it'll be coming in a few years anyway, why wait?
We found a way to inscribe data onto the face of individual leptons and quarks so pretty soon, 5 picometer chips!
TSMC probably has these nm's ready far in advance, they just release them incrementally to keep the sales going. Kind of like the diamond industry. I see right through this scheme.
I don't think many people here realise the magnitude of this technology. It's insane that we can create circuits this small, and they remain reliable and usable in the consumer world. Not even 10 years ago, this would have been seen as near impossible.
They are not baby steps. Going from 3nm to 2nm is a 33% improvement. So that is the same as going from 300nm to 200nm in one step.
1.4nm. 🤯
The stuff ASML and TSMC is able to pull off is peak human engineering.
Consider me seriously impressed.
The stuff ASML and TSMC is able to pull off is peak human engineering.
Consider me seriously impressed.
Pretty much I'm sure 17 Pro will be with N3P, than we will see N2. With yield issues and costs, they won't skip it.
You’ve got to look at the math. All of these are reducing the size around 30%. They have to move incrementally as they continue to develop the chips and the fabrication processes to make the chips
No way
They’ll have convinced people that 4GB is enough by then
Shared storage/ram
macOS will be a subscription cloud service
I'm not sure if this is satire or if you are serious? The manufacturing equipment and techniques for developing these advanced chips takes years or decades to develop. It's not like one day a TSMC engineer wakes up and says, I think I'll create a 10 Ångström chip today. The EUV machines from ASML cost over $150 million a piece and that is for 3 nm. Next generation is $300 million a piece.
Is your caps lock stuck? Also that has NOTHING to do with cooling, lower process size usually means the other, and due to the new process being more efficient they are able to get more performance at the same WATT output. So if you think for a second about it if they made the processors have the same performance for say the M3 VS the M2 they would be using less power. The chips are just more efficient so doing more work for the same budget.
It's the same naming format, it's the size and then the number for that size.
N2 is the same as A20.
A is for Angstrom that is 0.1 of a nanometre.
Also Apple just says NM size they don't reference TSMC process naming.
The "sub nm" labeled transistors are laboratory curiosities that can only be fabricated in microscopic quantities at enormous costs. The equipment to built that stuff in large quantities doesn't exist. Gazillions of dollars per year need to be spent over many years to gradually bring that cost down to something commercially viable. Cash flow limits how many tens or hundreds of Billions the industry can spend per annum. In the mean time, all those gazillions spent on fab equipment research and production make certain intermediate nodes become more affordable.
So at the current rate, TSMC will hit "zero" by 2050, and then in the last half of this century will see "negative nanometers".
Seriously, these numbers are just marketing names at this point. Intel had better names. Why can TSMC call these processes things like "stone lake" or "volcano tree"?
Seriously, these numbers are just marketing names at this point. Intel had better names. Why can TSMC call these processes things like "stone lake" or "volcano tree"?
I suppose Stone Age is good name for Intel.
Yeah I’ll wait for the 2nm. Should I get 17 Pro or wait a year for the better 2nm? But wait another year and have 1.4nm lol