Today in China Intel is holding their Technology and Manufacturing Day. Unlike previous "IDF" events this appears to be far more centered on the manufacturing aspects of Intel's latest process nodes. During presentations Intel talked about their latest steps down the process ladder to smaller and smaller geometries all the while improving performance and power efficiency.
Mark Bohr presenting at Intel Technology and Manufacturing Day in China. (Image courtesy of Intel Corporation)
It really does not seem as though 14nm has been around as long as it has, but the first Intel products based on that node were released in the 2nd half of 2014. Intel has since done further work on the process. Today the company talked about two other processes as well as products being made on these nodes.
The 10nm process has been in development for some time and we will not see products this year. Instead we will see two product cycles based on 14nm+ and 14nm++ parts. Intel did show off a wafer of 10nm Cannon Lake dies. Intel claims that their 10nm process is still around 3 years more advanced than the competition. Other foundry groups have announced and shown off 10nm parts, but overall transistor density and performance does not look to match what Intel has to offer.
We have often talked about the marketing names that these nodes have been given, and how often their actual specifications have not really lived up to the reality. Intel is not immune to this, but they are closer to describing these structures than the competition. Even though this gap does exist, competition is improving their products and offering compelling solutions at decent prices so that fabless semi firms can mostly keep up with Intel.
Nothing like handling a 10nm Cannon Lake wafer with bare hands! (Image courtesy of Intel Corporation)
A new and interesting process is being offered by intel in the form of 22FFL. This is an obviously larger process node, but it is highly optimized for low power operation with far better leakage characteristics than the previous 22nm FF process that Intel used all those years ago. This is aimed at the ultra-mobile devices with speeds above 2 GHz. This seems to be a response to other low power lines like the 22FDX product from GLOBALFOUNDRIES. Intel did not mention potential RF implementations which is something of great interest from those also looking at 22FDX.
Perhaps the biggest news that was released today is that of Intel Custom Foundry announcing and agreement with ARM to develop and implement those CPUs on the upcoming 10nm process. This can have a potentially huge impact depending on the amount of 10nm line space that Intel is willing to sell to ARM's partners as well as what timelines they are looking at to deliver products. ARM showed off a 10nm test wafer of Cortex-A75 CPUs. The company claims that they were able to design and implement these cores using industry standard design flows (automated place and route, rather than fully custom) and achieving performance in excess of 3 GHz.
Gus Yeung of ARM holding a 10nm Cortex-A75 based CPUs designed by Intel. (Image courtesy of Intel Corporation)
Intel continues to move forward and invest a tremendous amount of money in their process technology. They have the ability to continue at this rate far beyond that of other competitors. Typically the company does a lot of the heavy lifting with the tools partners, which then trickles down to the other manufacturers. This has allowed Intel to stay so far ahead of the competition, and with the introduction of 14nm+, 14nm++, and 10nm they will keep much of that lead. Now we must wait and see what kind of clockspeed and power performance we see from these new nodes and how well the competition can react and when.
If Gus is holding a wafer
If Gus is holding a wafer full of 10nm Cortex-A75 based chips, they must have a freaking ton of cores on them. Look at how large they are! Those are Xeon sized dies!
Well, good question. They are
Well, good question. They are test chips, so… extra logic? Lots of cores and caches? No real details were given other than it was a wafer of CPUs with A75 cores that were designed by Intel.
Looks like 9 die high and 11
Looks like 9 die high and 11 wide. So, given it’s a 300mm wafer, that’s 33.3 by 27.27 for a die area of 909mm^2 That’s safely larger than any reticle that I’m aware of!
“Gus Yeung of ARM holding a
“Gus Yeung of ARM holding a 10nm Cortex-A75 based CPUs designed by Intel. (Image courtesy of Intel Corporation)”
Did you mean Fabricated by Intel’s Chip fabs and not designed by Intel. Because the Cortex-A75 is an ARM Holdings reference design. So Intel had no part in that design other than on the Fabrication(Sub-Contractor) side.
And the A75 is still not a very wide order superscalar design compared to say Apple’s Custom A7 cyclone and newer designs that are twice as wide order superscalar as any of ARM holdings reference.
And What will be the Fate of Jim Keller’s Custom K12 ARMv8A ISA running CPU micro-arch that Keller/K12-Team designed for AMD at the same time Keller/Zen-Team was working their mojo on the reworked x86 ISA running Zen mico-arch. K12 and Zen are supposed to share the same DNA under the bonnet but it’s just that one desgn(Zen) is engineered to execute the x86 ISA and the other(K12) is engineered to run the ARMv8A ISA.
So K12 was moved back to 2018 while AMD focused on Getting Zen to market and now that that is complete the fate of K12 is still unknown! I think that Keller/Team’s custom K12 can give any Apple custom CPU/ARMv8A ISA running design some very good competition, but who knows when/if that will happen!
The wording that Intel sent
The wording that Intel sent out was that they designed the CPU… which features ARM cores and other interconnects. So the best guess here is that Intel synthesized the design themselves to work on their process to produce a final test product.
Well that sounds like ARM
Well that sounds like ARM holdings POP™ designs and I’ll bet that Intel was only the Fab partner with that Chinese interest doing the design with ARM’s standard POP™ IP blocks and ARM Holdings system integration services/help. I’ll bet that Arm Holdings helped more with the Chinese Client’s ARM IP system integration needs with Intel along mostly as the foundry sub-contractor.
And it would be big news if Intel was offering up system integration services using mostly ARM Holdings IP, other than Intel’s Fab services, for this client or any other potential clients.
ARM supply logical designs,
ARM supply logical designs, not physical layouts. If Intel wanted to do a ‘proof of concept’ for their entire design-to-tapeout service (automated layout etc) then Intel may have just done the ‘normal’ process of licensing the cores/interconnects and designed their own die. Those are some big dies, so they may just have packed everything they could into the reticle size as a “look how big we can go!” demonstration without any regard for practical functionality, as it’d not be a chip they intend to release for commercial sales. Plus, if they just picked as many of ARM’s available designs as possible to pack onto a die, that means they now have a library of ‘pre-debugged’ parts they can use if customers request those same parts again to save time & money.
Yes, right! In what
Yes, right! In what bizarro-world are Intel’s engineers going to just waltz right over Arm Holdings(Owned by Softbank/Partners, not a bank by the way) and get access to all the proprietary Arm Holdings circuit layouts for ARM Holdings’ reference design A75 core! Designs that require ARM Holdings’ engineers and ARM Holdings’ for a fee integration services. Do Tell me how Intel’s engineers could ever have the time to get up to speed on a design that Intel has not had the time to even look at without paying a large sum of money to ARM Holdings, Including a very limited number of Intel chip fab engineers being even allowed access to parts of the design and signing any NDA required.
Do you think that ARM Holdings is just going to let some Fab Partner have the Keys to the IP vault and even take that system POP(TM)/Other system Integration Business away from ARM holdings and all the revenues that that POP(TM)/other similar business produces for ARM Holdings above and beyond any chip IP licensing revenues/royalty revenues that ARM Holdings receives from its IP Licensees.
Intel is nothing more than a foundry sub-contractor in this deal and Intel does not have the IP license for that ARM Holdings POP(IP) above what is necessary to do the foundry part of the fabrication work only. ARM Holdings is the one with the reference design ARM core design/s IP, the on CPU die functional blocks IP(on die Interconnect, CPU core, GPU core IP, If the customer is using any Mali GPU IP), etc.
Intel is not doing that much work other than having its Intel chip fab engineers(Who are under an ARM Holdings Legal NDA, even to their higher ups at Intel) do some limited transistor design tweaking(Individual Transistor geometry and circuit doping regimens, etc.) to get the ARM Holdings layouts working with wherever Intel process node/doping recipes/layout libraries(Software defined circuit layout automated design parameters tweaking) that are being utilized for the fabrication of the Chinese customer’s product.
ARM Holdings does not directly produce the chips it designs but ARM Holding makes plenty of its revenues, above and beyond that Licensing and Royalty part of its revenues, providing system integration POP(TM)/other similar services for those customers that do not possess their own in-house chip design/chip system functional block design/integration engineers. So Unless more is Known about the deal Intel is most likely nothing more than a chip fab partner in the deal with ARM Holdings’ managing the project for its Chinese client.
Sure Intel is free to lease a Top Tier Architectural license(From ARM Holdings) for the ARM Holdings ARMv8A/Newer ISA and go about engineering its own custom CPU micro-arch to run that licensed ISA, that’s what Apple/AMD/Nvidia/Others do. But ARM Holdings is not very likely to give Intel any total access to any IP that is ARM Holdings’ bread and butter IP that ARM Holdings shares with no others. So Unless Intel has taken out a Top Tier Architectural license and is not using any ARM Holding’s reference design cores/other IP then Intel is most likely just a fab partner with very limited specific access for only a limited number of Intel’s chip fab/fab process engineers to any of that ARM Holdings IP.
“In what bizarro-world are
“In what bizarro-world are Intel’s engineers going to just waltz right over Arm Holdings(Owned by Softbank/Partners, not a bank by the way) and get access to all the proprietary Arm Holdings circuit layouts for ARM Holdings’ reference design A75 core! ”
By buying it. That is ARM’s core business model: selling logical designs to other companies who then implement and fab those designs. That’s what ARM /does/, it’s the only thing they do.
Arm Holdings does not sell
Arm Holdings does not sell any of Its IP outright and Arm Holdings only licenses the designs or licenses the ISA(Only) protion. Also ARM Holdings gets royalities on a per unit sold basis from some licensees. So stop all of your Daft replies and stop embarrassing yourself in the eyes of readers.
Apple Pays for a top tier architectural license on the ARMv8A ISA(Limited License) from Arm Holdings so Apple can use its own in house designs that are engineered by Apple’s engineers to run that licensed ARMv8A ISA but Apple does not Own the rights to the ARMv8A ISA as ARM Holdings retains thoes IP rights. The are many many ARM Holdings licensees with various levels of licensing agreements with ARM holdings and none of ARM holdings licensees has any unristricted rights or total ownership of any IP licensed from ARM Holdings.
You really are a simple minded egregious fool with your complete lack of Knowledge and Understanding of Arm Holdings’ business model and you are only pulling things out of your nether region at this point in time.
“Arm Holdings does not sell
“Arm Holdings does not sell any of Its IP outright and Arm Holdings only licenses the designs or licenses the ISA(Only) protion. ”
Nooope. ARM license out the core module designs, peripheral module designs, etc. Some companies customise those cores, some implement their own outright(e.g. Qualcomm’s Snapdragon), some just implement the stock ARM core (e.g. Mediatek). Why the heck do you think so many different companies happen to be using an ARM core design that just happens to be called ‘Cortex’?
” none of ARM holdings licensees has any unristricted rights or total ownership of any IP licensed from ARM Holdings.”
Well no shit Sherlock, who thinks they would? Other than you, clearly, as you’re the only one to have even implied it.
Intel licensing an ARM core design to fab on their process is no different than any other company licensing an ARM core design to fab.
Maybe a little AMD K12 with
Maybe a little AMD K12 with some AI IP for the Tesla Cars! Keller(Works at Tesla) must be leading that push!
“Tesla is working with AMD to develop its own A.I. chip for self-driving cars, says source”
https://www.cnbc.com/2017/09/20/tesla-building-an-ai-chip-for-its-cars-with-amd-globalfoundries.html
“This has allowed Intel to
“This has allowed Intel to stay so far ahead of the competition, and with the introduction of 14nm+, 14nm++, and 10nm they will keep much of that lead.”
Did Trump write this?