I just checked the GloFlo info and they're claiming the following. My thinking is that this would make an excellent process for mobile designs. Raven Ridge or follow-on would be unbeatable.
"If you look at performance with back-bias 22FDX is the same or better than 16/14nm FinFET process. With 12FDX with back bias you get better than 10nm FinFET processes"
In short 12FDX offers performance of 10nm FinFET with better power consumption and lower cost than 16nm FinFET.
Ok, but they're outright saying they're looking to produce completely different chips This is not me speculating, this is GlobalFoundries saying it themselves.
I think 12FDX would be great for the APUs, but I'm not sure its worth it for AMD to put the resources for a separate process for what are already low profit chips. And it appears that GlobalFoundries isn't terribly interested in even producing such chips. Now they probably will be producing ARM chips integrating modems (the mixed chips they were referencing). But their focus with their FDSOI processes seems to be a completely different class of chip.
It is relative to FinFETs built at GlobalFoundries only. GlobalFoundries by 2020 will have only one FinFET foundry, while two for FDSOI. (Officially)
-
-> The CTO stressed that the decision was made not based on technical issues that the company faced, but on a careful consideration of business opportunities the company had with its 7LP platform as well as financial concerns.
-> GlobalFoundries scuttles 7nm chip plans claiming no demand.
-> In announcing the move, Caulfield said companies don’t seem to have much interest in the planned 7nm architecture. Rather, they are planning to stay with the current-gen architectures and squeeze performance out by other means.
-> Gary Patton admits that GlobalFoundries never planned to be a leading producer of 7-nm chips in terms of volume. Furthermore, the company has been seeing increasing adoption of its 14LPP/12LP technologies by designers of various emerging devices, keeping Fab 8 busy and leaving fewer step-and-scan systems for 7LP products.
-> What is next:
1. Scaling Out the 14LPP/12LP
2. Investing in FD-SOI;
(quote)GF is intensifying investment in areas where it has clear differentiation and adds true value for clients, with an emphasis on delivering feature-rich offerings across its portfolio. This includes continued focus on its FDXTM platform, leading RF offerings (including RF SOI and high-performance SiGe), analog/mixed signal, and other technologies designed for a growing number of applications that require low power, real-time connectivity, and on-board intelligence. GF is uniquely positioned to serve this burgeoning market for “connected intelligence,” with strong demand in new areas such as autonomous driving, IoT and the global transition to 5G.(quote)
-
It does not say they are dropping leading edge on FDX. Since, the only comparison is Samsung currently. It is severely lacking in any competition, GlobalFoundries won before it even started.
FDSOI at GlobalFoundries before this decision was the sandbag child. GlobalFoundries didn't want FDSOI to compete with FinFETs. So, they bombed the benchmarks of FDSOI and removed any mention of high performance.
Only one foundry at GlobalFoundries is getting the GigaFab treatment:
https://www.anandtech.com/show/12534/change-of-strategy-globalfoundries-3-0/3
The most expensive UTBB FDSOI is still cheaper than the least expensive FinFET. When GlobalFoundries does not have their own product on the benchmark:
http://www.chinaisgood.com/wn/14/zcixagxa.html
22FDX obliterates the 14nm-based 22FFL node from Intel.
You're intentionally ignoring what I'm pointing out (this has been a recurring problem for you). Its not about the process at all (I'm not even criticizing the process, just your insistence about your speculation being fact), its about GlobalFoundries saying they don't plan on making those types of chips using that process (you ignore that again in your own post, see point 2 where they explicitly say again they're targeting their FD-SOI processes for RF and wireless chips).
Yes, it doesn't explicitly say that, but it sure sounds to me like GF is largely abandoning their development of new processes (not just FinFET ones, and FD-SOI has been behind FinFET - they're still developing 12, which while they claim it can compete with 10nm FF, FF is at 7nm now - so leading edge chips are not going to be looking to use it), or at least pushing their development out significantly, which means they're probably going to be years behind and so any company trying to offer a leading edge product is not going to use them. Like you said, they're leading edge in FDSOI by default right now because Samsung is the only other one even bothering to push on it and they're not exactly pushing hard. But after 12FDX they haven't said they're going to be pushing hard on furthering FDSOI. And they're coming right out and saying they're targeting other types of chips with their current 22 and new 12 processes. I don't think there's any reason AMD couldn't make chips using it, but it doesn't sound like they're planning to and doesn't sound like GF is counting on them to. And it doesn't sound like anyone else making leading edge chips are either. There for sure will be some ARM based chips but those will likely be lower end ones or specialized ones and not cutting edge ones (which I don't think there's an inherent problem with the process preventing that, just that because ARM targeted the development of their newer designs for the leading FF process at the time, that it'd take effort - $ - to change it for FDSOI so most just aren't going to).
Yes, we know FDSOI is cheaper, that's why its viable for lower cost chips. It does not offer the best performance for leading edge processors. That is why they won't be made using it.
I have to wonder though, if GF might just sit 7nm idle (not fully abandon things just yet, but halt their own development), let others develop it and then just license their production later. Like if 5nm development struggles (which seems probable), extending 7nm life. I would've thought if nothing else GF go to TSMC and Samsung and see if they want to lease or buy the 7nm facilities for their production. They'd take over development, but would get extra facilities. And with all the talks about tariffs, it might prove beneficial to have some production elsewhere.
No doubt, but that's a prioritization issue, not a process maturity issue. And given the demand for 7nm, TSMC has an incentive to maximize production on this lucrative process as much as they can.
I wasn't remarking on process maturity at all, and absolutely. The issue is, they have finite amount of 7nm production capacity (its not like they can just build a new 7nm fab within a couple of months if they need more capacity). This is a market that can't react that quickly to demand. And if they mistakenly expect more production than they actually receive they could end up very bad off (probably what is happening to GF, where because of their failures with 14nm and 28nm, that a lot of customers pivoted away, and then they saw they weren't going to have enough customers for 7nm so they'd end up losing a lot of money by developing it for hardly anyone).
I am am pretty shocked by this, this is terrible for AMD it really is.
What is zen 2 going to be fabbed on?? Is there going to be a delay? Capacity shortages?..aweful.
What I can say is well done to Lisa su for having some contingency plan, wonder if AMD will hit Samsung up for their 7nm process?..they lost Qualcomm to TSMC so room for AMD to step in?...how long would that take to move over all everything?
Without capacity there is no chance AMD can shift the units required to claw back market share..even if they produce an awesome product, this can't be a good situation for AMD.
Hopefully glofo was honest with AMD all the way through and gave them a long heads up before this, and play fair with the agreement, let's be honest Glofo have royally screwed AMD over down the years with their incompetence!.
Pretty shocking, i was looking forward to this process, had potential to be highest clocking.
Intel sure are laughing right now, they are back in the game.
Edit; AMD have confirmed to Forbes that it is using TSMC for all of it's upcoming processor's and multiple tapeouts have taken place with TSMC, no setbacks are predicted.
https://www.forbes.com/sites/marcoc...sors-to-be-manufactured-at-tsmc/#5407070b4642 Looks like AMD had a heads up or Lisa Su is a fortune teller.
I see you saw the new info.
Either GF was pretty honest with AMD, or AMD could pretty easily tell how much GF was struggling with 7nm, which is why Zen 2 is at TSMC. AMD knew GF would not be producing on time. I don't think they expected them to quit 7nm though, and were likely expecting to be able to use it in the future (hoping they'd need to even, by having product in high enough demand that they'd need the extra production capacity). Short term, I'd guess this actually benefits AMD. Since they knew GF was behind TSMC they decided to get going with the latter big time. And they probably got a better price than they will after this news where TSMC will almost certainly raise prices (although, now that AMD will need more production there, that will be higher priced probably - although it might be offset by reduction in the WSA penalty). With AMD's product stack, along with their partners (Sony and Microsoft), I'm not sure if that has benefits (like it costs less money to produce things that are similar to what you've already produced there), that might give them a legup on Nvidia. But with Nvidia's big chips, that also creates a problem for them, so this actually is potentially a pretty big break for AMD's GPU division.
Did Qualcomm move back to TSMC? I thought they had moved a lot to Samsung? Or is Samsung's 7nm not as far along?
But the two big questions now are, how good is TSMC 7nm, and how/when is Samsung's 7nm? (Ok that's more than 2 questions)
Could we be have this reversed?
It appears to the public, that GloFlo was moving ahead full speed until this announcement. Is it possible that AMD decided to abandon GloFlo for 7nm as they were having delays. I think that AMD was their lead/only customer for 7nm and thus loosing them resulted in no demand for the process.
Possibly, although I don't think its quite so simple. I do believe AMD was still intending to use GF 7nm (they seemed somewhat caught off guard by this, saying they had a product looking to be taped out at GF in Q4 of this year), but GF was looking and realized AMD and IBM were their only 7nm customers, and for the cost of development and overcoming the problems (either for yields or it wouldn't be competitive in performance, or some manner of both) would require so much money and likely not get much in return (as AMD showed they're not going to just wait on GF). I can see where GF is coming from as they have just been failing. Its sad that 7nm didn't work out though, as it seemed like it had potential to get them more on the right track, but apparently that wasn't actually the case.
You're the one that fails to see how the Apple / TSMC economic works for others. Apple funds the new process investments for TSMC in advance for exclusive production during the first two quarters. Volume production ramp up starting around july, building inventory until iphone launch in september, peak sales of iphone last until the new year. From january on, sales of iphone wind down and capacity is freed up. From jan/feb AMD can tape out their 7nm Zen 2 and have it on the shelves around may. AMD / nVidia and some other large customers have the advantage they can follow in Apple's slipstream.
No, I don't struggle seeing that. The thing is, with most of the industry looking at TSMC, they have finite 7nm fabs. They can't just build a new one quickly to meet demand. Apple isn't paying TSMC to build out more fab space (as Apple doesn't need it and sure doesn't want to subsidize competitors; they just want priority). So everyone will be waiting for Apple to get priority (so they'll already be behind and waiting; yes it helps some in that Apple helps spur development and get things through the early phase where the process isn't as mature so certainly its not all bad, but its not like they can magically make process development go faster or improve either so any problem and these other companies could be looking at an extended delay on when they can start production), and then they will have to try and work deals where they speculate on how much production they need with little room to adjust from there (and now with no GF, TSMC can jack up the price with less competition). And if TSMC itself has issues, they will be limited in what they can do about it unless Samsung is more competitive. And if Apple has a product that is selling hot or they expect it to, they can gobble up more of the capacity for longer time. Heck, if Apple just decides that they want to "compete" by just hogging as much capacity as possible, making it so that other products can't get produced, they could. They have the money to be able to lock up TSMC production and create headaches. So far, Apple hasn't had a reason to do that, so its been mostly fine.
This isn't the end of the world, but there's definitely potential for problems. Outright, I think we'll be looking at higher prices though, which some of that was just that it takes a lot of resources for making leading edge chips on leading edge process, and that has been getting more complex and costly, but couple that with reduced competition and its gonna push prices up.
There is one potentially good aspect, I think this might speed up the work on chiplets. But there could be some rough years ahead as companies struggle with developing that alongside trying to develop new process tech.
Facebook
Twitter