Lunar Lake test CPU benchmarks show up on SiSoft

Very cool. Do we have any tables with feature sizing density for intel 18A vs competition? This is going to be using all-around gates and backside power delivery right?
 
serpretetsky said:
Very cool. Do we have any tables with feature sizing density for intel 18A vs competition? This is going to be using all-around gates and backside power delivery right?
Click to expand...
https://www.tomshardware.com/news/intel-completes-development-of-18a-20a-nodes

Intel's 20A fabrication technology will rely on gate-all-around RibbonFET transistors and will use backside power delivery. Shrinking metal pitches, introducing all-new transistor structures and adding backside power delivery at the same time is a risky move, but it is expected that 20A will allow Intel to leapfrog the company's competitors — TSMC and Samsung Foundry. Intel plans to start using this node in the first half of 2024.

Intel's 18A manufacturing process will further refine the company's RibbonFET and PowerVia technologies, as well as shrink transistor sizes. The development of this node is apparently going so well that Intel pulled in its introduction from 2025 to the second half of 2024.
 
As much as I dread an Intel return to dominance, somebody needs to compete with TSMC. Samsung isn't there yet, but they're working on it. 3 cutting-edge foundry options would be a huge improvement.

I have to wonder, though, does anybody see a future where AMD uses Intel foundries to make their chips? Now that they're past 14nm+++++ and their advanced nodes are lining up, what are the chances that Intel spins off its foundries?
 
I would not dread an intel return to dominance vs TSMC&the rest, even would it happen, the Apple, Microsoft, Google, Facebook, Tesla, Nvidia, Amazon are so ridiculous a strong client-need-demand for cutting edge foundry to make things happen and will work hard for their fate to not be in a single entity hands/single point of failure too much, and demand being so strong for the second best options to still find clients in the near future to be kept alive even if someone has a clear lead.

It would be pure superbe news, if Intel top node > TSMC top node in 2024 because of an intel success and not a TSMC giant failure.
 
Last edited:
jfreund said:
As much as I dread an Intel return to dominance, somebody needs to compete with TSMC. Samsung isn't there yet, but they're working on it. 3 cutting-edge foundry options would be a huge improvement.

I have to wonder, though, does anybody see a future where AMD uses Intel foundries to make their chips? Now that they're past 14nm+++++ and their advanced nodes are lining up, what are the chances that Intel spins off its foundries?
Click to expand...
What’s funny is Intel managed to get 18A working on the Twinscan NXE scanners with the .33NA optics instead of the .55 ones, Intel won’t have delivery of those .55 units until 2025 so they already have their 18A+ node lined up.

Higher the NA the better the resolution.

So Intel has a lot of fun things in the works, I don’t think AMD would move to Intel for foundry work any time soon though. They would need a complete redesign and it would be an image thing, they would go Samsung before they went Intel.

Though seeing an AMD processor and an Intel one in an Apples to Apples comparison where they have the same node, and same power targets for a head to head architecture comparison would be cool, I doubt it would happen.

But Intel already has spun off its foundries.
https://www.extremetech.com/computi...manufacturing-groups-into-a-separate-business
 
Last edited:
Does anyone else think Intel's cycles are becoming just a bit too short? I mean I had a slightly defective motherboard that was only corrected later with a BIOS flash. I ended up thinking I'd need to wait 9 months after a new platform comes out before buying in the first place.
 
wandplus said:
Does anyone else think Intel's cycles are becoming just a bit too short? I mean I had a slightly defective motherboard that was only corrected later with a BIOS flash. I ended up thinking I'd need to wait 9 months after a new platform comes out before buying in the first place.
Click to expand...
Yes. I saw something yesterday about a supposed lunar lake refresh. Lunar lake is already 3 "generations" out and we're talking about a refresh? Refreshing is already what Intel does when they're not ready with a real next gen so if they already think they need to do that 3 years from now, they need a reality check. Admittedly, that rumor could just be bs.

What's extra annoying is where I'm thinking about ultrasonic from Alder lake to Raptor refresh, but meteor and arrow are supposedly coming with maybe big improvements. It's enough to make you want to just day the bell with it and hold off for a few more years (which, yeah I know a bunch of you do). To be fair Alder was a huge upgrade from Kaby Lake.
 
Intel got its dick kicked in by AMD, TSMC, and ARM.
Their refresh cycle rate is advancing about as fast as new nodes opening to them can accommodate at a price that is reasonable.
Intel 7 is not on par with TSMC 7 it’s better than TSMC 10 but it’s not good enough.
Intel 4 is good but it took too long to bring online, they don’t have enough manufacturing capacity to supply a mainstream major launch and TSMC 4 is easier for them to access but it’s a minor embarrassment for Intel to be paying top dollar to their major competitors to use their process.
Intel 20 and 18 are their first chance in a long time to say FU we’re #1 and the mobile space is heating up, Apple is gaining share in enterprise because their devices are better pound for pound than anything Dell, Lenovo, or HP is offering in that space.
Intel and their shareholders are very tired of seeing their dominant position chipped away as their products are mocked and surpassed. They have bruised egos and rather than make excuses they’ve come out swinging.
It’s a good time.
 
serpretetsky said:
The foundry is still part of intel right? Just a different business group. Unlike the amd spinoff global foundries.
Click to expand...

Right. Intel spun-off/broke-out its foundries from the core, not sold/got rid of them. Separate bookkeeping, money still ends up in Intel's pocket.
 
Last edited:
You must log in or register to reply here.
Back
Top