Monday, August 22nd 2022

TSMC (Not Intel) Makes the Vast Majority of Logic Tiles on Intel "Meteor Lake" MCM

Intel's next-generation "Meteor Lake" processor is the first mass-production client processor to embody the company's IDM 2.0 manufacturing strategy—one of building processors with multiple logic tiles interconnected with Foveros and a base-tile (essentially an interposer). Each tile is built on a silicon fabrication process most suitable to it, so that the most advanced node could be reserved for the component that benefits from it the most. For example, while you need the SIMD components of the iGPU to be built on an advanced low-power node, you don't need its display controller and media engine to, and these could be relegated to a tile built on a less advanced node. This way Intel is able to maximize its use of wafers for the most advanced nodes in a graded fashion.

Japanese tech publication PC Watch has annotated the "Meteor Lake" SoC, and points out that the vast majority of the chip's tiles and logic die-area is manufactured on TSMC nodes. The MCM consists of four logic tiles—the CPU tile, the Graphics tile, the SoC tile, and the I/O tile. The four sit on a base tile that facilitates extreme-density microscopic wiring interconnecting the logic tiles. The base tile is built on the 22 nm HKMG silicon fabrication node. This tile lacks any logic, and only serves to interconnect the tiles. Intel has an active 22 nm node, and decided it has the right density for the job.
The CPU tile is the only logic tile built on an Intel node, which in this case is the Intel 4 node. The company considers this process to be on-par or better than TSMC's N5, and it probably wanted the crown jewels of its IP—CPU cores—to be built on a native fab. The CPU tile contains the CPU cores, a last-level cache, and Foveros interfaces.

The Graphics tile is the second-most important logic tile, and contains an iGPU based on the Xe-LPG graphics architecture. An evolution of Xe-LP, the LPG features real-time ray tracing capabilities. Intel decided to use the TSMC N5 (5 nm EUV) node for this tile. Not all of the iGPU is based on this tile, some of it, such as the Display Engine, could be located on the I/O tile.

The SoC tile is the largest in terms of area, and is built on the TSMC N6 (6 nm) node. This contains the memory controllers, PCIe root-complex, and the controllers and SerDes (serializer-deserializer) of the various on-package devices. The I/O die is the smallest die, and is essentially an extension of the SoC die. It's built on the same TSMC N6 node, and features the PHY (physical layer) components of the various I/O.
Source: PC Watch
Add your own comment

41 Comments on TSMC (Not Intel) Makes the Vast Majority of Logic Tiles on Intel "Meteor Lake" MCM

#1
ZoneDymo
herpa derp "glue" egg on face, alright, did the obligatory (and totally deserved) shaming, moving on.
Posted on Reply
#2
john_
Considering Intel's manufacturing, that's probably good. But then again, they might face the same capacity constrain AMD is facing. Because I believe, if AMD had Intel's capacity, it could be selling double the number of CPUs and GPUs it sells today. And I don't expect Intel to have payed TSMC for "infinite" capacity.
Posted on Reply
#3
phanbuey
The goal here is to use apple/samsung/ibm etc tiles.

Their whole platform is designed to just use whatever the best is at the current moment, so this makes total sense. This is kind of cool.
Posted on Reply
#4
Dirt Chip
I like the idea of different tiles, will see how it goes. Many supply and assembly problems can make it a long wait but any how I'm upgrading this round so I don't really care.
Posted on Reply
#5
Nanochip
Tomshardware is reporting that intel will begin using UCIe to connect the chiplets/tiles in the future, starting with the generation after Arrow Lake.
Posted on Reply
#6
ppn
That means the Desktop could get integrated PCH in the shape of IO or SOC tile. Finally no more PCH on the motherboard that is $30-50 of the cost reduced. why would 1800 have 100 more pins if not for the PCIe + sata that the PCH is responsible for.

The IO tile could be the integrated MC+ DDR5 Phy since it's directly glued to the Compute tile. Can't imagine IMC in the SOC it's too far could cause latencies..

Besides the left and right images show the CPu rotated differently along the short side. so it's very consfusing. If they want to keep adding cores. only the right image makes sense.
Posted on Reply
#7
AM4isGOD
Just like they invented 3D cache on AMD CPU's, poking fun at Intel works both ways

Meteor Lake is the CPU that "could" win back the crown for Intel, interesting times ahead.

Foveros is really interesting too, if anyone cares to read up on it.
Posted on Reply
#8
caroline!
ppnThat means the Desktop could get integrated PCH in the shape of IO or SOC tile. Finally no more PCH on the motherboard that is $30-50 of the cost reduced.
Companies will still charge the same to consumers though, that's how capitalism works.
Posted on Reply
#9
Wirko
ppnThat means the Desktop could get integrated PCH in the shape of IO or SOC tile. Finally no more PCH on the motherboard that is $30-50 of the cost reduced. why would 1800 have 100 more pins if not for the PCIe + sata that the PCH is responsible for.
Even with existing Intel CPUs it would be possible to build a zero-chipset low-end system, basically analogous to (barely existing) AMD A300 or X300. The only thing a mobo maker would have to add is USB and networking. And, ugh, the Management Engine.
ppnThe IO tile could be the integrated MC+ DDR5 Phy since it's directly glued to the Compute tile. Can't imagine IMC in the SOC it's too far could cause latencies..
But the iGPU needs a high bandwidth path to memory too, which would likely overload the compute tile's interfaces.
ppnBesides the left and right images show the CPu rotated differently along the short side. so it's very consfusing. If they want to keep adding cores. only the right image makes sense.
Seeing such a small tile with the cores and a giant SoC tile, I dare to speculate that the last level cache is part of the Soc tile. This would be beneficial to Intel in several ways. Firstly, the LLC would serve the graphics part too, like it does in existing Intel chips. Secondly, cache shrinks far less than logic when going to a finer node. This also works the other way: a static RAM cell can't be much larger on TSMC N6 than it is on Intel 4. If EMIB works as advertised (able to join the four Sapphire Rapids tiles practically into one, needing minimum or zero interface logic) then the interface between the cores and LLC can have low enough latency despite sending data between tiles.
But there's at least one downside: if they're going to have variants of compute tile with more and fewer cores, it would make sense for the LLC to scale up and down too, which can only be done if they have a variants of SoC tile as well.

What I don't dare to speculate - just thinking aloud - is that some cores will also find their place on the Soc tile. P or E, I have no idea which ones would fit N6 better. SoC tile is just ... too big. Are there any estimates of tile sizes out there?
caroline!Companies will still charge the same to consumers though, that's how capitalism works.
Capitalism of old: company innovates, finds a way to reduce cost without reducing value to consumer. [I don't care if there's just a small rock and three wires inside my PC, as long as it works as a PC.]
Capitalism evolved: you get less for more.
Posted on Reply
#10
Gica
john_Considering Intel's manufacturing, that's probably good. But then again, they might face the same capacity constrain AMD is facing. Because I believe, if AMD had Intel's capacity, it could be selling double the number of CPUs and GPUs it sells today. And I don't expect Intel to have payed TSMC for "infinite" capacity.
Everything that is high tech in the processor is manufactured at Intel.
Stop crying on AMD's shoulder. They assassinated customers' wallets in the last two years and the least important ones, home users, were practically ignored.
Posted on Reply
#11
Jimmy_
caroline!Companies will still charge the same to consumers though, that's how capitalism works.
i think their Desktop SKUs won't have this I/O tile in them.
Posted on Reply
#12
john_
GicaStop crying on AMD's shoulder.
New here? From WCCFTECH?
GicaEverything that is high tech in the processor is manufactured at Intel.
Try making that CPU without the TSMC parts. We are not talking about a car where someone else makes the Tires. "Just buy the car and buy tires latter from someone else". Doesn't work like that.
GicaThey assassinated customers' wallets in the last two years and the least important ones, home users, were practically ignored.
Assassinated customer wallets. Yeap. WCCFTECH....
Posted on Reply
#13
ppn
WirkoWhat I don't dare to speculate - just thinking aloud - is that some cores will also find their place on the Soc tile. P or E, I have no idea which ones would fit N6 better. SoC tile is just ... too big. Are there any estimates of tile sizes out there?


Capitalism of old: company innovates, finds a way to reduce cost without reducing value to consumer. [I don't care if there's just a small rock and three wires inside my PC, as long as it works as a PC.]
Capitalism evolved: you get less for more.
The soc is estimated to 95 mm² by SemiAnalysis. Still bigger than the Soc part in alder lake that is only 84 if we take into account the media engine system agent and PCie, ddr5 phy, display ports and so on. So the PCH is a mistery, maybe the IO itself. But too small to be that.

Previously was supposed to be on 5nm, now it turns out to be 6nm, postponed to lunar and nova lakes.

intel did cut out the CPU part out of Raptor lake, shrinked it to 5/4nm, but that didn't result that much of a smaller die, just 10-12 mm² smaller than Alder lake that has less L2, and even less than Raptor.

I see now, the first image shows the 2+8 tile, the IO is shorter. Probably limited to PCIe 4x, the desktop version should be 16X +4 for nvme
2+8, 6+8 and 8+16 tile are 37,5, 68 and 105 mm² info in this thread.

www.techpowerup.com/294799/intel-meteor-lake-2p-8e-silicon-annotated
Posted on Reply
#14
enzolt
Never understood why computer nerds either bend over for Gelsinger or bend over for Su. Vote with your wallet. Nobody but yourself cares if you have an AMD or Intel. Build your PC and go watch your porn.
Posted on Reply
#15
AM4isGOD
enzoltNever understood why computer nerds either bend over for Gelsinger or bend over for Su. Vote with your wallet. Nobody but yourself cares if you have an AMD or Intel. Build your PC and go watch your porn.
Too much red care on TPU, blue is brown.

Makes zero difference who makes what for either Intel or AMD as long as they are good. Don't see why it matters unless you have some vested interest in them.
Posted on Reply
#16
Gica
john_Try making that CPU without the TSMC parts. We are not talking about a car where someone else makes the Tires. "Just buy the car and buy tires latter from someone else". Doesn't work like that.
Tell me which part of Intel processors (4004-13th ... 1971-2022) comes from TSMC? If Taiwan collapses, Intel are the only ones capable of manufacturing top processors. Now they cooperate with TSMC just to maximize profit, not that they can't manufacture them.
Posted on Reply
#17
john_
GicaTell me which part of Intel processors (4004-13th ... 1971-2022) comes from TSMC? If Taiwan collapses, Intel are the only ones capable of manufacturing top processors. Now they cooperate with TSMC just to maximize profit, not that they can't manufacture them.
I can't take seriously what you post. Really. I just can't. You throw stuff, just to throw stuff, stuff that are meaningless. We are talking about a future processor that integrates parts of TSMC and you talk about 4004. You talk about Taiwan's collapse, which really it's not even an argument here, it's pointless. I can only understand this part of your post as wishful thinking of an Intel shareholder. Shares will skyrocket. But in this thread, it's a pointless comment. I wouldn't call it an argument. And then you finish your post with a little love for Intel. "They can build it, it's just cheaper to go to TSMC". The same love you showed in your last post. "Hi end parts are made at Intel, the rest, not so hi end, at TSMC". My God...... The iGPU is made in the TSMC. Is it low end tech? I mean. Even YOU can debunk your own posts.

I am not going to reply to you again. I only replied to this post of yours, because I just couldn't believe what I was reading.
Posted on Reply
#18
Gica
It's very easy to understand if you don't see everything in red.
Since 1971, they have been manufacturing their own processors, they don't need anyone, but this policy will change. Not that they don't want to, but because it's more profitable for them.
As for Taiwan, it will be very difficult for them to produce anything if China invades the island. The scenario is not impossible and in the event of a war, Intel remains the only ones capable of manufacturing top processors. The others will be forced to adapt to Samsung's limited production or turn to Intel factories.
john_The iGPU is made in the TSMC. Is it low end tech? I mean. Even YOU can debunk your own posts.
Yes, it is not high end. It's just a...igpu. The computing units, the cores, are the well-guarded secrets.
Posted on Reply
#19
AM4isGOD
GicaIt's very easy to understand if you don't see everything in red.
Since 1971, they have been manufacturing their own processors, they don't need anyone, but this policy will change. Not that they don't want to, but because it's more profitable for them.
As for Taiwan, it will be very difficult for them to produce anything if China invades the island. The scenario is not impossible and in the event of a war, Intel remains the only ones capable of manufacturing top processors. The others will be forced to adapt to Samsung's limited production or turn to Intel factories.


Yes, it is not high end. It's just a...igpu. The computing units, the cores, are the well-guarded secrets.
However much people here might poke fun at Intel, they have their own FABS, AMD do not, which is a definite advantage for Intel. They also have enough spare cash to buy production capacity from TSMC, win in both ways. At some point AMD will lose out by not having their own FAB. However badly Intel might be doing with their processes at the moment, they have the means and the capacity to improve them without having to rely on anyone else, which at some point will pay off. Once they get it/them right, they just have to ramp up the production and roll the cash in.

How can anyone see it as a downer Intel buying production from TSMC? Would you not do the same if you could if it gives you a advantage.
Posted on Reply
#20
Al Chafai
Things are going to get interesting with all these 3D packaging techniques, we are in for a treat boys, the next decade is gonna be awesome for tech enthusiasts and general consumers.
Posted on Reply
#21
ncrs
ppnThe IO tile could be the integrated MC+ DDR5 Phy since it's directly glued to the Compute tile. Can't imagine IMC in the SOC it's too far could cause latencies..
Looks like it is in the SOC tile:

It's hard to see due to low resolution, but there are "Memory Control" and "Memory IO" present on the right side. The slide is from their official presentation at Hot Chips.
Posted on Reply
#22
Tatty_Two
Gone Fishing
Cleaned up some, let's try to talk a little more about the topic and a little less geopolitics ..... thank you.
Posted on Reply
#23
Gica
Al ChafaiThings are going to get interesting with all these 3D packaging techniques, we are in for a treat boys, the next decade is gonna be awesome for tech enthusiasts and general consumers.
It is the next logical step. Horizontally they reached the maximum.
Intel's collaboration with TSMC is based on the principle: in factory X I can produce an indispensable processor part for $5, and I don't have space to produce one for $10. Wouldn't it be good for others to manufacture the $5 part for me and for me to produce more $10 units?
Posted on Reply
#24
mahirzukic2
GicaIt is the next logical step. Horizontally they reached the maximum.
Intel's collaboration with TSMC is based on the principle: in factory X I can produce an indispensable processor part for $5, and I don't have space to produce one for $10. Wouldn't it be good for others to manufacture the $5 part for me and for me to produce more $10 units?
That could be one way of looking at it.
Another one would be, what if I make this 5$ part some place else, exactly the place where my main competitor produces their stuff, and directly cutting available capacity for my main competitor.

Remember that the last 2 quarters AMD couldn't produce enough chips (both cpu and gpu) for the market, everything they produced, they sold right away.
Imagine being in a situation that you can't produce stuff fast enough (as it is flying off the shelves like hot cookies) and some other company comes in and takes some capacity off of your supplier. That's what's going on here. That is the primary reason, the (cheaper) cost of producing it on TSMC is secondary to this notion.
Posted on Reply
#25
Gica
Contracts at this level have nothing in common with those on the street corner. They are perfected years before the product launch.
AMD is dependent on TSMC, Intel is not. It is not difficult to guess who has priority and who does not.
It's pure speculation with AMD sales. All released Zen 3 processors were available on the market and had a high reference price (msrp) from the start.
Posted on Reply
Add your own comment
Apr 18th, 2024 07:44 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts