• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

TSMC Ships its 1 Billionth 7nm Chip

btarunr

Editor & Senior Moderator
Staff member
Joined
Oct 9, 2007
Messages
47,670 (7.43/day)
Location
Dublin, Ireland
System Name RBMK-1000
Processor AMD Ryzen 7 5700G
Motherboard Gigabyte B550 AORUS Elite V2
Cooling DeepCool Gammax L240 V2
Memory 2x 16GB DDR4-3200
Video Card(s) Galax RTX 4070 Ti EX
Storage Samsung 990 1TB
Display(s) BenQ 1440p 60 Hz 27-inch
Case Corsair Carbide 100R
Audio Device(s) ASUS SupremeFX S1220A
Power Supply Cooler Master MWE Gold 650W
Mouse ASUS ROG Strix Impact
Keyboard Gamdias Hermes E2
Software Windows 11 Pro
In a bid to show off its volume production prowess and technological edge (but mostly to rub it in to rival fabs), TSMC on Thursday announced that it shipped its 1 billionth chip fabricated on its 7 nm process. If these dies were combined into one big rectangular wafer, they would cover 13 New York City blocks. TSMC's 7 nm process debuted with its N7 node, which went into volume production in April 2018, over two years ago. The fab has since mass-produced 7 nm chips for the likes of Qualcomm, Apple, and AMD, among dozens of other clients. The company now looks to monetize refinements of N7, namely the N7e and N7P (DUV refinements), while executing its crucial EUV-based N7+ node, leading up to future nodelets such as N6. Much of TSMC's growth will be propelled by 5G modems, application processors, and its pivotal role in the growth of companies such as AMD.



View at TechPowerUp Main Site
 
Regardless of whether or not these 7 nm chips are equivalent to the Intel 10 nm process, the fact of the matter is every one is on a better process than all 14 nm chips shipped by Intel. And TSMC shipped a billion of them. And with only an extremely small number of Intel chips shipped on a process below 14 nm, TSMC is way, way ahead.
 
Regardless of whether or not these 7 nm chips are equivalent to the Intel 10 nm process, the fact of the matter is every one is on a better process than all 14 nm chips shipped by Intel. And TSMC shipped a billion of them. And with only an extremely small number of Intel chips shipped on a process below 14 nm, TSMC is way, way ahead.
I don't care what the Intel guys say, 14 NM is old tech..it gets really tiring to see these guys trying to polish a turd. Reminds me of a used car salesman.
 
I don't care what the Intel guys say, 14 NM is old tech..it gets really tiring to see these guys trying to polish a turd. Reminds me of a used car salesman.

Ok here is me adding a little tiny bit to that.. and I'm a guy that runs 2 Ryzen 3000 rigs and looking to go all AMD when RDNA2 comes out because i'm a huge fan of what AMD is doing atm... BUT.. HOWEVER... Intel.. Old tech YES.. but no.. their 14nm is still competing head to head with AMDs 7nm... And yeah marginally but having a slight edge is more CPU (single core) intensive games... Obviously the gap is closing rapidly now and they will be wiped of the face of the planet with Zen 3 for sure... but you have to admire how much they have been able to squeeze out of that almost ancient node now... everything they will develop after this will pale in comparison from that perspective.

And no I'm also not investing in technology from 2015 when I can choose way more efficient tech from 2020 at better prices... they are being silly with their pricing! but development wise. Damn they did a stellar job at squeezing that 14nm!
 
Ok here is me adding a little tiny bit to that.. and I'm a guy that runs 2 Ryzen 3000 rigs and looking to go all AMD when RDNA2 comes out because i'm a huge fan of what AMD is doing atm... BUT.. HOWEVER... Intel.. Old tech YES.. but no.. their 14nm is still competing head to head with AMDs 7nm... And yeah marginally but having a slight edge is more CPU (single core) intensive games... Obviously the gap is closing rapidly now and they will be wiped of the face of the planet with Zen 3 for sure... but you have to admire how much they have been able to squeeze out of that almost ancient node now... everything they will develop after this will pale in comparison from that perspective.

And no I'm also not investing in technology from 2015 when I can choose way more efficient tech from 2020 at better prices... they are being silly with their pricing! but development wise. Damn they did a stellar job at squeezing that 14nm!

Exactly. Skylake uArch and 14nm node stretched all the way over 5-6years that's really an insane feat in engineering no doubt at all, 7nm AMD bleeding edge + Archtecture refreshes from Zen, Zen+ to Zen 2 and now Zen 3 vs that 14nm++. I also really doubt Intel can deliver such technology again, I think all those minds left Intel or dunno what the hell is wrong with them, BK was fired on the pretext of some metoo garbage rather than his incompetence and now they are more into M&A and pursuing external fab for their Xe GPU division to save more money and face, that Bob Swan needs to go back to his CFO and Intel should get a new person who can steer their ship to proper path.
 
And here I thought football fields (not sure if US or proper football fields though, hence why hockey rinks would be easier), buckets or bananas were the commonly used measurements for comparison, not Manhattan city blocks (are they bigger or smaller than blocks in other cities?). I'm also curious how many non Americans even know what a city block is...
 
To me, it simply boils down to this:

Intel = 14nm milk upon 14nm milk upon moar 14nm milk, still equals... wait for it...........................fattier 14nm milk, as in their bank accounts....
AMD = 7/6/5/4/3/2/1/0mn...constantly pushing tech progress fowards... like it should be
 
And here I thought football fields (not sure if US or proper football fields though, hence why hockey rinks would be easier), buckets or bananas were the commonly used measurements for comparison, not Manhattan city blocks (are they bigger or smaller than blocks in other cities?). I'm also curious how many non Americans even know what a city block is...
13 NYC blocks = 20.8 acres. At 96.5 MTr/mm2 for N7, it works out to 8.12 quintillion (8.12*10^18) transistors.
 
13 NYC blocks = 20.8 acres. At 96.5 MTr/mm2 for N7, it works out to 8.12 quintillion (8.12*10^18) transistors.
Can you point out those city blocks on a map?
Also, what's an acre? Is that like a jia? I mean, TSMC are Taiwanese after all...
 
Last edited:
Can you point out those city blocks on a map?
Also, what's an acre? Is that like a jia? I mean, TSMC are Taiwanese after all...
Okay, so 13 NYC blocks is roughly 84,200 square meters.
 
I do confess, upon reading the article, I did get a cuckle out of that "13 New York City blocks" unit of measurement. How many Chevy Tahoes is that? ( :
 
The constant argument that intelv10nm is equal to TSMC 7nm get old. Either way, they have been in volume production for over 2 years. Ice lake production is tiny, but even using that as volume production for 10nm, it was still a year and a half later, while TSMC is ramping 5nm right now with its first high volume production chipset being the A14, and Intel justannounceda delayto its 7nm node. Intel will learn from this and eventually catch up or be "close enough", but it's going to probably take a decade.
 
Manhattan city blocks (are they bigger or smaller than blocks in other cities?). I'm also curious how many non Americans even know what a city block is...
Good question, being an Aussie I'm aware of what city blocks are, but never thought to question if there is a global standard size.
 
Back
Top