G.Skill TridentZ 5 RGB 6800 MHz CL34 DDR5 review
Be Quiet! Dark Power 13 - 1000W PSU Review
Palit GeForce RTX 4080 GamingPRO OC review
Core i9 13900K DDR5 7200 MHz (+memory scaling) review
Seasonic Prime Titanium TX-1300 (1300W PSU) review
F1 2022: PC graphics performance benchmark review
MSI Clutch GM31 Lightweight (+Wireless) mice review
AMD Ryzen 9 7900 processor review
AMD Ryzen 7 7700 processor review
AMD Ryzen 5 7600 processor review
Apple and Intel will be the first customers for TSMC's new 2 nm manufacturing node.
TSMC's new N2 node will be used by Apple and Intel, according to reports and sources in the financial world. N2 is TSMC's first manufacturing process to use the GAAFET (Gate-All-Around Field-Effect Transistor) design. By the end of 2025, it should be in wide use.
TSMC will be late to the GAAFET party. Samsung's 3GAE node will come out in 2023, and Intel's first Angstrom-era process, Intel 20A, will come out in 2024.
"We also see more clarity around TSMC's N2 expansion schedule in Fab 20 (Hsinchu)," Sze Ho Ng, an analyst with China Renaissance Securities, wrote in a note for clients. "Tool move-in is expected to start by end-2022, based on company plans, ahead of risk production in late 2024E with Intel (client PC Lunar Lake's graphic 'tiles', while the CPU 'tiles' are fabbed using Intel’s 18A) and Apple being the anchor customers for dedicated capacity support."
Intel's own slide that talks about the graphics tiles of Meteor Lake, Arrow Lake, and Lunar Lake processors clearly says that the latter's GPU will be made outside of the computer. This is because Intel's own technology is better than N3.
AMD, Broadcom, Nvidia, and MediaTek have all said that they will use different nodes from TSMC's N5 family (N5, N5P, N4, N4P, N4X). MediaTek has already formally announced its Dimensity 8000/8100 application processors and Dimensity 9000 SoC. Nvidia, on the other hand, will use a custom 4N fabrication process for its Hopper and presumably Ada Lovelace GPUs, which will be made in a different way. Also, AMD will make its Genoa and Raphael processors with a 5 nm process.
In a report from DigiTimes, all of these companies are in talks with TSMC about getting N3-capable capacity from the company in late 2020 or early 2024. In addition, these companies are also expected to start talking about N2-capable allocations next year. Apple and Intel, on the other hand, will be the first to use N2.
First, TSMC will use gate-all-around field-effect transistors (GAAFET) in its N2 technology. This is years after Samsung's 3GAE and more than a year after Intel's 20A technology, which both used this type of technology (2024). In terms of power, performance, and area/transistor density, the world's largest contract chip maker hasn't said what to expect from N2 over N3. However, because this will be a new node, it is reasonable to think that it will be better than its predecessors. The new process will still use the same extreme ultraviolet (EUV) lithography scanners that have a 0.33 numerical aperture. Intel's 18A, on the other hand, is set to use ASML's Twinscan EXE EUV scanners with High-NA (0.55NA).
Intel EVP is optimistic about 2021 - not concerned about competition from Apple and AMD - 12/10/2020 09:57 AM
Intel Executive Vice President Gregory Bryant conversed with market analysts from Swiss bank UBS and Yahoo! Finance. Here sales figures for the past year have been discussed, they also had an outlook ...
Battle Royale Inbetween Apple and EPIC - to terminate dev accounts that affects the Unreal Engine - 08/18/2020 09:18 AM
Apple threatens to remove EPIC from the developer program. Earlier on w reported that Apple single sidely removed Fornite from the Apply Store distribution platform. ...
Apple announces Mac transition to Apple silicon - 06/23/2020 08:11 AM
Apple today announced it will transition the Mac to its world-class custom silicon to deliver industry-leading performance and powerful new technologies. Developers can now get started updating their ...
Apple and Broadcom have to pay university 1 billion after patent infringement - 01/30/2020 10:24 AM
Apple and the American chip maker Broadcom must jointly pay an amount of 1.1 billion dollars (1 billion euros) to the California Institute of Technology (Caltech)....
Apple Announces Apple Arcade - 03/26/2019 08:53 AM
Apple today Apple Arcade, a game subscription service that will feature over 100 new and exclusive games, including original releases from renowned creators Hironobu Sakaguchi, Ken Wong, Will Wright a...
Maddness
Senior Member
Posts: 2188
Joined: 2014-10-24
Senior Member
Posts: 2188
Joined: 2014-10-24
#6011985 Posted on: 04/25/2022 10:06 AM
The big problem with Global Foundries was they weren't competitive with cutting edge technology. It may well have just come down to revenue. But I'm glad they still produce some stuff for AMD. But I'm very glad they moved onto TSMC.
The big problem with Global Foundries was they weren't competitive with cutting edge technology. It may well have just come down to revenue. But I'm glad they still produce some stuff for AMD. But I'm very glad they moved onto TSMC.
schmidtbag
Senior Member
Posts: 7143
Joined: 2012-11-10
Senior Member
Posts: 7143
Joined: 2012-11-10
#6012063 Posted on: 04/25/2022 02:59 PM
I've always said that this is just a sneaky way of Intel being anti-competitive. They're not doing anything illegal and technically, you could argue they need to do this if their own nodes aren't doing what they want/need. But, it either makes AMD's contracts more expensive or limits how many chips AMD can get from TSMC. Even if Intel sells fewer chips than AMD, just by simply making it harder to buy AMD chips is a win for Intel.
If AMD kept GF, I think they'd be in much worse shape than they're in today.
For one thing, GF would have been underfunded pre-2017, where they wouldn't have even got 12nm. So, while 1st-gen Ryzen would have been cheaper for AMD to produce, it would have been worse simply because the node was less mature.
AMD would have put some money into GF though, and that would have meant less money going toward Ryzen. So, the architecture itself would have been worse too.
I think AMD is at a point where they could re-merge with GF, but they're so technologically behind that they would likely struggle just to get to 7nm, let alone 2nm. AMD would be better off just starting from scratch.
Intel has its own fabrication plants pouring out ~100% Intel chips, and in addition it will claim a portion of TSMC's production capability. That's one smart, if not underhanded, way of beating the competition.
I've always said that this is just a sneaky way of Intel being anti-competitive. They're not doing anything illegal and technically, you could argue they need to do this if their own nodes aren't doing what they want/need. But, it either makes AMD's contracts more expensive or limits how many chips AMD can get from TSMC. Even if Intel sells fewer chips than AMD, just by simply making it harder to buy AMD chips is a win for Intel.
I wonder if the folks at AMD ever regret letting go of GF and consequently GF dropping out of the cutting edge technology race entirely, ending up as a nameless contract fab instead. Although if splitting off GF saved AMD from bankruptcy, I reckon regretting it wouldn't be relevant.
If AMD kept GF, I think they'd be in much worse shape than they're in today.
For one thing, GF would have been underfunded pre-2017, where they wouldn't have even got 12nm. So, while 1st-gen Ryzen would have been cheaper for AMD to produce, it would have been worse simply because the node was less mature.
AMD would have put some money into GF though, and that would have meant less money going toward Ryzen. So, the architecture itself would have been worse too.
I think AMD is at a point where they could re-merge with GF, but they're so technologically behind that they would likely struggle just to get to 7nm, let alone 2nm. AMD would be better off just starting from scratch.
H83
Senior Member
Posts: 4471
Joined: 2009-09-08
Senior Member
Posts: 4471
Joined: 2009-09-08
#6012066 Posted on: 04/25/2022 03:04 PM
I also dislike this move by Intel because it´s going to hurt AMD and Nvidia more than doing any favours to Intel himself...
But this dick move is not illegal so Intel can go ahead and cripple competition even before they get their products out of the factory...
As for AMD still owning GF instead of selling, that would be even worse because AMD didn´t have the money to invest in GF so they could invest the same money on better nodes.
Selling GF was a life saver for AMD!
Intel has its own fabrication plants pouring out ~100% Intel chips, and in addition it will claim a portion of TSMC's production capability. That's one smart, if not underhanded, way of beating the competition. I wonder if the folks at AMD ever regret letting go of GF and consequently GF dropping out of the cutting edge technology race entirely, ending up as a nameless contract fab instead. Although if splitting off GF saved AMD from bankruptcy, I reckon regretting it wouldn't be relevant.
I also dislike this move by Intel because it´s going to hurt AMD and Nvidia more than doing any favours to Intel himself...
But this dick move is not illegal so Intel can go ahead and cripple competition even before they get their products out of the factory...
As for AMD still owning GF instead of selling, that would be even worse because AMD didn´t have the money to invest in GF so they could invest the same money on better nodes.
Selling GF was a life saver for AMD!
tunejunky
Senior Member
Posts: 3011
Joined: 2017-08-18
Senior Member
Posts: 3011
Joined: 2017-08-18
#6012095 Posted on: 04/25/2022 04:18 PM
H83 & Schmidtbag are correct.
silicon fabrication is expensive & difficult.
selling GF was a force multiplier for AMD as it's stock is over five times higher (than the date of sale), it's market share is more than doubled, and it's technology leadership is unquestioned (at least in uArch).
compare and contrast that to AMD being an afterthought
yes Intel is a huge swollen dick, yes they are being anti-competitive, but what else is new?
the funny thing about this is it is a double edged sword for both Intel and TSMC.
for Intel it means temporarily edging out AMD @ 2n. - but AMD is not as vulnerable to "process wars" as Intel since they've gone to chiplet based uArch. they can do for one model year what Intel has done for ten (re:uArch).
for TSMC it means dealing with the devil because this fat contract allows long term competitiveness - by underwriting new fabs in America. for Intel it's a stop-gap to allow their fabs to catch up - but once they do "it's on like Donkey Kong" and we'll see the real fab wars.
what i'm hoping for is Samsung to improve their density and yields to truly become a large scale player in contract fabbing instead of the third rate player they are now.
H83 & Schmidtbag are correct.
silicon fabrication is expensive & difficult.
selling GF was a force multiplier for AMD as it's stock is over five times higher (than the date of sale), it's market share is more than doubled, and it's technology leadership is unquestioned (at least in uArch).
compare and contrast that to AMD being an afterthought
yes Intel is a huge swollen dick, yes they are being anti-competitive, but what else is new?
the funny thing about this is it is a double edged sword for both Intel and TSMC.
for Intel it means temporarily edging out AMD @ 2n. - but AMD is not as vulnerable to "process wars" as Intel since they've gone to chiplet based uArch. they can do for one model year what Intel has done for ten (re:uArch).
for TSMC it means dealing with the devil because this fat contract allows long term competitiveness - by underwriting new fabs in America. for Intel it's a stop-gap to allow their fabs to catch up - but once they do "it's on like Donkey Kong" and we'll see the real fab wars.
what i'm hoping for is Samsung to improve their density and yields to truly become a large scale player in contract fabbing instead of the third rate player they are now.
Click here to post a comment for this news story on the message forum.
Senior Member
Posts: 3287
Joined: 2013-03-10
Intel has its own fabrication plants pouring out ~100% Intel chips, and in addition it will claim a portion of TSMC's production capability. That's one smart, if not underhanded, way of beating the competition. I wonder if the folks at AMD ever regret letting go of GF and consequently GF dropping out of the cutting edge technology race entirely, ending up as a nameless contract fab instead. Although if splitting off GF saved AMD from bankruptcy, I reckon regretting it wouldn't be relevant.