Tuesday, March 2nd 2021

TSMC to Start 3 nm Node Production This Year

Taiwan Semiconductor Manufacturing Company (TSMC), the leading provider of semiconductors, is supposed to start 3 nm node production this year. While Samsung, one of the top three leading semiconductor foundries, has been struggling with the pandemic and delayed its 3 nm node for 2022, TSMC has managed to deliver it this year. According to a report, the Taiwanese semiconductor giant is preparing the 3 nm node for the second half of this year, with the correct date of high-volume product unknown. The expected wafer capacity for the new node is supposed to be around 30,000 wafers per month, with capacity expansion expected to hit around 105,000 wafers per month in 2023. This is similar to 5 nm's current numbers of 105,000 wafers per month output, which was 90,000 just a few months ago in Q4 2020. One of the biggest customers of the upcoming 3 nm node is Apple.
Source: Hardwareluxx.de
Add your own comment

7 Comments on TSMC to Start 3 nm Node Production This Year

#1
voltage
I wonder if leakage will go up again? (cross talk etc.) or will 3nm have better control with less leakage? If it has more again, as the last previous die shrinks, that's not a good sign.
Posted on Reply
#2
mtcn77
voltageI wonder if leakage will go up again? (cross talk etc.) or will 3nm have better control with less leakage? If it has more again, as the last previous die shrinks, that's not a good sign.
I think it is up to the designers how close they want the transistors. It can be dense and slow, or fast but not breaking the trend...

You get one of these: "either 40% dense, 25% less power, or 15 faster". The designers make it for you which is pretty important decision. Technically it is all that seperates sequential cpu generations.
Posted on Reply
#3
bonehead123
GO TSMC .. :respect:..:clap:.. :)

T he
S martest
M anufacturing
C ompany
Posted on Reply
#4
ppn
mtcn77You get one of these: "either 40% dense, 25% less power, or 15 faster"
It's never the case that you get the same size and less power at same clocks, but rather smaller die, same power and more frequency. So we do get all of those at the same time. When they say 40% denser they always mean compared to previous node. never between the Performance and low power variants of the same node. If you meant 40% denser compared to the same node that would explain alot why AMD 7nm is the same exact density as 10nm in every aspect. Nvidia 7nm 65,6, AMD 7nm 41 is in fact 40% less dense, so what kind of trickery is that.
Posted on Reply
#5
AleksandarK
voltageI wonder if leakage will go up again? (cross talk etc.) or will 3nm have better control with less leakage? If it has more again, as the last previous die shrinks, that's not a good sign.
Leakage is always going up whenever you shrink a node. That is why there is stuff like GAAFET(that Samsung uses) to better control it. Don't know what TSMC is using, however, they are still managing to control it. A lot of things also depend on the design you are trying to manufacture.
Posted on Reply
#6
mtcn77
ppnIt's never the case that you get the same size and less power at same clocks, but rather smaller die, same power and more frequency. So we do get all of those at the same time. When they say 40% denser they always mean compared to previous node. never between the Performance and low power variants of the same node. If you meant 40% denser compared to the same node that would explain alot why AMD 7nm is the same exact density as 10nm in every aspect. Nvidia 7nm 65,6, AMD 7nm 41 is in fact 40% less dense, so what kind of trickery is that.
Thanks, but I'll hold on to my views.

You may go after the elusive performance target, but it is a stupid target in gpus, imho. There aren't any gpu architectures that push power without a tdp constraint, so power becomes more important than performance, imho.

I think AMD keeps it constant because it is a new microarchitecture and it is more efficient than the previous, thus omitting the density cost that a previous microarchitecture would entail. Nvidia does what Nvidia can. They always try to perform spectacularly, though not without a patchy streak.
Posted on Reply
#7
pjl321
Would 30,000 be enough for the A15 chip this year or not even close?
Posted on Reply
Copyright © 2004-2021 www.techpowerup.com. All rights reserved.
All trademarks used are properties of their respective owners.