2nm and 1.4nm are already on the roadmap. Expect them around 2024, 2026.
And 1nm is in the work if you are wondering.
Last edited:
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.
TSMC to Begin Production of 3nm Chips This Year
- Thread starter MacRumors
- Start date
- Sort by reaction score
And 1nm is in the work if you are wondering.
But they "are" factually behind. Doesn't matter how you slice and dice it.
How much behind is up to debate though. If you account for volume and shipment, my opinion is about 3 years.
Remember 3nm doesn't necessarily actually mean 3nm, it's a bit more complicated than that in practice. It is clear TSMC are now out ahead for transistor density though, and that allows for some superlative chips!
It is more a marketing trick than anything else. Nothing in this chip is 3nm small but, hey it sounds cool and helps the shares go up!
A claim is one thing to be doing it is one thing, but shipping in volume is the other. Just ask Intel about the difference.
Minimum 2yr lead on the rest of the industry. Superlative is appropriate at a minimum.
So what is the equivalence figure? Is Intel's 10nm equivalent to TSMC's 7nm or 5nm?There used to be a 'standard' but now each manufacturers process is different. It has been proposed to now measure density, in which case Intel could potentially 'win' as their 10nm process has more density than TSMC's 7nm process.
A Density Metric for Semiconductor Technology [Point of View]
Since its inception, the semiconductor industry has used a physical dimension (the minimum gate length of a transistor) as a means to gauge continuous technology advancement. This metric is all but obsolete today. As a replacement, we propose a density metric, which aims to capture how advances...ieeexplore.ieee.org
Or, to put it another way, if Intel were measured in the way TSMC's are measured, what would the figure be?
(somehow, I think I'm going to be told it isn't possible to do it this way!!
)What is with this attitude of everyone should just buy everything? Stop....
Taiwan based TSMC and Dutch based ASML are working on a 2nm-node production proces. TSMC uses for 3nm still FinFET transistors but for 2nm the GAAFET will be used. ASML is the supplier of EUV photolithography systems producing the chip wafers. ASML in collaboration with Belgian IMEC is developEd 1nm (1.5nm at first) lithography technology with commercialization of new photolithography system expected in 2022. But the first systems are just for early test production and chips with 2nm or 1.5nm will be some years away.
Roadmap doesn’t mean theyll ever see the light of day. Could face major production issues down the line.
Yes but isn’t this really about the technique and tech needed to write on silicone at that size at all? Is there any reasons AMD couldn’t have a desktop made from this?
The A16 chip will most likely be based on this 3nm tech, not the 4nm one. Historically it wouldn’t make sense either. 4nm would probably be a backup design in case something went wrong with 3nm. (You can use your same 5nm design on the 4nm tech but it isn’t much better.)
After reading that post in the Intel thread, about how die size isn't what people think, I wonder if this is just so much PR-speak. Not blaming them for doing it if this was their goal, but it's kinda shady if so. But if 'fudging' in press releases is to be outlawed, wouldn't most of the ads running now be in violation? Hmm...
Nothing is impossible, just lack of technology. Once, people thought that reaching the moon would be impossible...
I can't wait for the point of "Hi... Apple. We would like to put your "Nvidia card" into our next chip set. Like.. no physical card.. just part of the GPU.
Technology will come full circle, they will be going back to vacuum tubes after 3nm.
Negative, space/particles/size gets smaller, not negative.
1 nm is 1000 picometer so a next step might be 0.9 nm=900 picometer.
Last edited:
Never heard of a risk cpu, do you mean RISC?
I remember back in 1995, while finishing my degree, taking a class on VLSI, where we were talking about how the design rules for chips were going to change significantly when we went below 0.3 µm (300 nm), because propagation delay along wires would be greater than propagation delay through logic gates. Up until then, the assumption was that wiring delay was nil and gate delay was all that mattered. Copper interconnects were still a few years off.
25 years later and we've almost moved the decimal two places.
The physical limit for electrical separation has a lot of factors going into it, but for reference the gate insulator layer (silicon dioxide insulator isolating the transistor gate from the channel) is about 3-5 atoms thick.
Also bear in mind that a transistor is much larger than the process feature size, though the latter drives the former. so "3 nm" transistors may well be spaced 50 nm apart (guessing there). You need room for the wiring, too, which is also driven by feature size.
25 years later and we've almost moved the decimal two places.
The physical limit for electrical separation has a lot of factors going into it, but for reference the gate insulator layer (silicon dioxide insulator isolating the transistor gate from the channel) is about 3-5 atoms thick.
Also bear in mind that a transistor is much larger than the process feature size, though the latter drives the former. so "3 nm" transistors may well be spaced 50 nm apart (guessing there). You need room for the wiring, too, which is also driven by feature size.
"Risk production" means "production to test the actual yield output of the process," i.e. you're debugging the tools and procedures used to make chips. The chips themselves are likely just arrays of test transistors, wired in a matrix so they can be individually tested and statistics can be gathered for how many work, sorta work, and fail. These statistics will be used to debug and improve the process, as well as the computer models used to design and validate the chips.
Assuming that all goes well, they're still months from cranking out "engineering sample" processors. If things don't go well, they're even more months behind as they need to go back and redesign.
In short, "what are the business risks in this process?" Not to be confused with RISC.
Assuming that all goes well, they're still months from cranking out "engineering sample" processors. If things don't go well, they're even more months behind as they need to go back and redesign.
In short, "what are the business risks in this process?" Not to be confused with RISC.
It’s just mind blowing that humans can make this stuff. We’re talking about the wavelength of X-rays.