News TSMC's N2 process has a major advantage over Intel's 18A: SRAM density

Status
Not open for further replies.
It just demonstrates the uselessness of comparing node names. "NM" means nothing and has meant nothing since the turn of the century. Even at the height of everyone shouting that TSMC''s nodes were "more advanced", Intel's "10nm" node was the equivalent of TSMC's "7nm" node. (There were larger and smaller objects for each like this article describes.)

TSMC took one look at Intel 14a on their roadmap a couple of years ago and simply changed their own roadmap to indicate their 1.5 node was now 1.3 with no physical changes so it seemed to be more advanced.

Process node names are no longer indicative of anything but the enthusiasm of the respective companies' marketing departments and should be relied upon for nothing.

As far as this article goes, don't you think the text in the first comment above would have been a better headline? It certainly seems more neutral as you report in the article that we really don't know yet how this will shake out.
 
So about +19%, and in the area that Intel could have 30 MiB, TSMC could deliver about 36 MiB. It's not nothing.

AMD might choose to go from 32 to 36 MiB if they switch to a (standard) 12-core chiplet for Zen 6. That would be a light, easy option instead of 48 MiB. That's expected to use N3 though.
 
  • Like
Reactions: coolitic
After Pat Gelsinger's eviction, I doubt 18A will even pan out. 10% yield is devastatingly low for 18A. They need to outsource 18A Panther Lake entirely to TSMC now.
 
What about Intel 20A SRAM density? You complete glossed over that. It's also GAA.
Does it matter, 20A is dead, officially cancelled and was always goiong to be a short-lived stopgap anyway, it's not being used in any consumer products at all.

As for 18A having BSPD, TSMC has their own solution coming which is actually more advanced than BSPD as used by Intel but will be a few years behind, probably for A16 in late 2026.
 
  • Like
Reactions: shady28
If Intel is actually able to get 18A into a shipping product in 2025 or early 2026, it would give them the first 2nm class x86 CPUs given that the first 1-2 years of TSMC N2 will likely be monopolized by Apple A19 and M5.
 
If Intel is actually able to get 18A into a shipping product in 2025 or early 2026, it would give them the first 2nm class x86 CPUs given that the first 1-2 years of TSMC N2 will likely be monopolized by Apple A19 and M5.
Intel already signaled that 30% of 18A Panther Lake will be outsourced/manufactured by TSMC. I guarantee that will be increased to 100% very soon.
 
Intel has never said that 18A equals 18 angstroms. It's just a marketing name to hide the fact that they're lagging behind in terms of engraving finesse. 18A = 3nm.

"Intel's 18A manufacturing process (1.8nm-class)", not true
 
Last edited:
  • Like
Reactions: phead128
Is SRAM density really a process thing, or a design thing, or both?
Do we really know that TSM can produce this super density and that it works?
One of the big problems with all these new nodes is even when you can produce dense features it can end up running hot, and the easiest solution can be to just thin it out. Maybe TSM's is going to need heroic heat management or water cooling.
 
18a got one major advantage -made in USA , TSMC 2nm is made in Taiwan. As you know Donald Trump........
18A is 3nm, it's not a competitor to TSMC's 2nm. That would be 14A which is non-existent.


Also, Intel has signalled that 30% of 18A will be outsourced to Taiwan TSMC, and after Pat's departure, I guarantee that will rise to 100% soon.

TSMC has said they will move 2nm to Arizona by 2026 or 2027. So your point is moot and the national security angle is highly overplayed by people who don't know Semiconductors at all.
 
Intel already signaled that 30% of 18A Panther Lake will be outsourced/manufactured by TSMC. I guarantee that will be increased to 100% very soon.
I have no idea where you think you got that information. Perhaps Intel is still planning on using TSMC to build the graphics chiplet and you consider that "30%"? Assuming Intel executes according to plan TSMC will not even be capable of making 18A for another year or more as it's all based on ASML's High NA machines that TSMC doesn't yet have installed (though should get delivered this quarter I think.)

18A is 3nm, it's not a competitor to TSMC's 2nm. That would be 14A which is non-existent.

Take a deep breath and repeat after me, NM means nothing, NM means nothing. Go back to my original reply because NM means nothing and has meant nothing since, oh 300 nm just before the turn of the century. Since then it's just been marketing jargon. Most any expert will tell you that 18A will be a more advanced node than anything TSMC has if Intel executes it as planned. But, there is absolutely no relationship between "18A" and the size of any feature on the chip. It's just marketing at this point and you're caught up in it.

The reason articles like this are written is because it's useful to compare actual features on the chips because the actual process name is useless and has been for decades.

Here's a similar article from a couple of years ago comparing TSMC' s "7nm" process to Intel's "10nm" process. Yeah, not much difference there. 18A, 2.0c, Tastes Great, Less Filling, Built Ford Tough, No One Out-Pizza's the Hut, all of it is just marketing jargon and has no relationship to reality. No angstroms will be harmed in the making of Intel's 18A chips and I can get a much better pizza here than Pizza Hut!
 
I have no idea where you think you got that information. Perhaps Intel is still planning on using TSMC to build the graphics chiplet and you consider that "30%"?
It's reported by Tom's hardware in August. Yes, Panther Lake 18A processors are partially manufactured at TSMC. While the majority (approximately 70%) of the silicon area for Panther Lake is being produced in-house by Intel using its 18A process technology, the remaining 30% is manufactured externally by TSMC.

https://www.tomshardware.com/pc-com...intel-fabs-nova-lake-almost-entirely-in-house
Assuming Intel executes according to plan TSMC will not even be capable of making 18A for another year or more as it's all based on ASML's High NA machines that TSMC doesn't yet have installed (though should get delivered this quarter I think.)
18A is equivalent to 3nm in performance, which TSMC can manufacture currently using low-NA EUV for N3 (3nm nodes). You don't need high-NA EUV to make 3nm (18A). Also, TSMC will be at HVM with 2nm with low-NA EUV well before 18A 3nm gets tape-out in mid-2025. High-NA EUV is a mere optical (lens) optimization, it's not a game-changer like DUV vs. EUV.
 
Status
Not open for further replies.