Question [AT] AMD extends GloFo WSA to 2025

[NostaSeronx]

If AMD will guarantee hundreds of millions of dollars to continue producing 14/12nm parts, you can be damn sure those lines will keep running.

There are factors leading AMD becoming #2 at GloFo or lower. So, this idea of AMD saving FinFETs at GlobalFoundries is pretty absurd. Since, it damn sure didn't keep 7nm running.

There is multi-billions for 45nm PDSOI/FDSOI and 22/12 FDSOI, versus just one customers hundred millions?

 

[NTMBK]

There are factors leading AMD becoming #2 at GloFo or lower. So, this idea of AMD saving FinFETs at GlobalFoundries is pretty absurd. Since, it damn sure didn't keep 7nm running.

There is multi-billions for 45nm PDSOI/FDSOI and 22/12 FDSOI, versus just one customers hundred millions?

See the edit on my last post But more to the point- 7nm required a lot more R&D and investment in equipment. Whereas 14nm and 12nm are installed, running, and profitable. If AMD will guarantee enough wafer purchases to keep those production lines full until 2025, then they will keep running. That is pure profit for GloFo.

 

[NostaSeronx]

Whereas 14nm and 12nm are installed, running, and profitable.

No, GlobalFoundries 14nm/12nm FinFETs is sold at a loss.

If AMD will guarantee enough wafer purchases to keep those production lines full until 2025, then they will keep running. That is pure profit for GloFo.

Pure loss for GlobalFoundries for the one major reason of 45nm SOI being so big as it is at Fab8. SOI is profitable where FinFETs are not.

The adoption of 14LPP was meant to attract customers from Samsung. Which went nowhere since Samsung did 14LPC/14LPU/11LPP and GlobalFoundries did 12LP/12LP+. As is 14LPP/12LP/12LP+ has the lowest figures of merit.

Fab8 does not have anything competing against Intel Oregon's 22FFL or Samsung's Austin 17LPV/18FDS. It also doesn't provide better options than 14GP-Intel and 11LPP-Samsung.

There is also the removal of exclusivity for 14nm/12nm FinFETs. Which means AMD does not have to use GlobalFoundries for N-1 FinFETs products. The outcome is that AMD is making a move away from GlobalFoundries.

TSMC 16nm/12nm is the new home for Ryzen/EPYC for N-1 going forward. Not AMD committing more money for N-1 at GlobalFoundries.

AMD will move to the tune of GlobalFoundries or be trimmed out. Nothing is stopping GlobalFoundries in making a better business investment. Remember the pivot cut off five customers who were pathed to 7nm. Which AMD was #1 customer of at first.

GlobalFoundries moves Malta's 32nm/28nm capacity to Dresden.
GlobalFoundries runs purely 14nm/12nm at Malta.
First year = Negative net income || Dresden&Singapore Positive Revenue / Malta Negative Revenue
Second year = Negative net income || Dresden&Singapore Positive Revenue / Malta Negative Revenue
Third year = Negative net income || Dresden&Singapore Positive Revenue / Malta Negative Revenue
Fourth year = Negative net income || Dresden&Singapore Positive Revenue / Malta Negative Revenue
Fifth year = Negative net income || Dresden&Singapore Positive Revenue / Malta Negative Revenue
GlobalFoundries starts "GlobalFoundries will transfer its ongoing operations from Fab 10, as its East Fishkill fab is called, to its other facilities, including the massive Fab 8 plant in Malta."

Sixth year = Lower negative net income || Dresden&Singapore Positive Revenue / Malta Lower Negative Revenue
Seventh year = Brief positive net income || Dresden&Singapore Positive Revenue / Malta Brief Positive Revenue (45nm SOI surpassed 14nm/12nm FinFET wafer starts)
Then, there is the FDX node family also taking space of Fab8 eventually:

In turn converting Malta from big loss-zilla to big gains-kong.
For 2023:
12FDX/22FDX same tools used --{ 14LPP/12LP/12LP+ }-- 45SOI same cleanroom space
With very little to no capacity planned for FinFETs as they get squeezed by demand from tens of billions multi-customer rather than hundred of millions single-customer. In this case, the wants of the many outweigh the wants of the one. However, I am not the one thinking AMD's only purchase going forward is only 14nm/12nm FinFETs at GlobalFoundries.


We can use UMC's revenue as a guide for GlobalFoundries profitability, since their footprint is the same.
Negative revenue to income contribution by 14nm/12nm.
Positive revenue to income contribution by 28nm/22nm.
Positive revenue to income contribution by 55nm/45nm/40nm.
etc.

Jason Wang of UMC => "In addition, we expect the adoption rate of our 22nm technologies will continue to gain traction" - Q2 2021
"Revenue from 28-nanometer technologies continue to rise while business engagement in 22-nanometer have led to a growing number of customer tape-outs across wireless, display and IoT markets, further diversifying our product pipeline." - Q3 2021
Do to 28nm demand the price went from $1600 per wafer to $2300 per wafer.
2021-05-11 "The 28nm process price per wafer is approximately US$1,800, an increase of nearly 13% from the US$1,600 in the second quarter. In addition, industry insiders said that UMC will raise prices again in the first quarter of next year. By then, the price of 28nm process wafers will soar to US$2,300 per piece, an increase of nearly 28% from US$1,800."
2021-12-25 "For the same 28nm process technology, TSMC's previous quotation was US$2400-2800; for comparison, UMC's quotation was US$2100-2200, GlobalFoundries' quotation was US$2100-2300, and SMIC's quotation was US$1800-1900. Based on the current TSMC 28nm process quotation, it is estimated that the highest price is about US$2,800. After a 20% increase, it will be higher than UMC's 28nm process technology in January 2022 to nearly US$3,000."
GloFo's 28nm is 2100-2300 thus 1.95x 14nm193i = $4100-4485

The several customers for FinFETs(2020 Annual Report) at UMC have yet to appear. However, the more than 10 customers(2020 Annual Report) for 22nm have definitely appeared.
UMC = $627 million net income, Q3'21
GloFo = $5 million net income, Q3'21

This is reflected by GlobalFoundries getting design prototype wins for 14LPP/12LP/12LP+ but the design production is done at Samsung or TSMC instead. In this case, AMD is transferring cIOD/sIOD/Raven2/Raven/Summit/etc dies over from GloFo to TSMC as soon as possible.

There is no further adoption of 14nm/12nm FinFETs to Dresden or Singapore. However, there is adoption indications of 22FDX at Malta and Singapore. Thus reducing 22FDX load of, NA+LA markets with Malta and GC+APJ markets with Singapore. In turn giving more 22FDX to each major region more than one Fab in one region can.


The Singapore option is the most important since the extremely low-cost Turion/Sempron processors were fabbed over there.


Going further some analyst groups in 2021 have noted 28ULP(28SHP/28A rebrand) and 22FDX is getting prepped at Fab8.
The 28SHP/A to 28ULP renaming follows UMC's weird DDC node names 28HPM -> 28HPC -> 28HPCu+ -> 22ULP.

AMD's May 2015 Low-cost cancellation is dependent on increased costs from 20nm -> 14nm -> 7nm -> 5nm. While, Low-cost by AMD coming back is based off 28nm -> 22nm FDSOI -> 12nm FDSOI which have decreased costs going forward.

Margin's decrease 20nm -> 14nm -> 7nm -> 5nm; Low-cost is unsustainable
Margin's increase 28nm -> 22nm -> 12nm -> 6nm (7FD CEA-Leti/IMEC/GF); Low-cost is sustainable

AMD May 2015 = Low-cost is dropped
GloFo July 2015 = 22FDX is announced
AMD 1H2016-2H2018 = Ultra-Low-Power Architecture for CPU&GPU
GloFo September 2016 = 12FDX is announced
CEA-Leti and #1 Foundry Partner 2020/2021 = 7FD(44CPP/36Mx) is announced in the works

Dali Mobile = January 2020, Dali Desktop = July 2020, Pollock Mobile = July 2020
28nm EOL = January 2021 LTB(Last Purchase), July 2021 LTS(Last Recieved)

Mendocino Mobile = 2022, Mendocino Desktop = 2022 (killing off demand of 3000G(OEM)/3050GE(OEM))
Raven2/Dali >$50 -> Mendocino >$100
Stoney/Carrizo-L >$25 -> 22FDX >$25

Dali/Pollock >$50 (100% performance at Iso-freq/150 mm2) to 22FDX >$25(90% performance at Iso-Freq/90 mm2)
Stoney/Carrizo-L >$25 (~70% performance at Iso-freq/107-125 mm2) to above.

CPU/GPU is not as critical when the most important aspect is the Multimedia core.
14LPP/12LP+ Multimedia core
Reducing leakage by Vt = RVT -> HVT -> sHVT -> eHVT (four options; HVT onwards costs process steps)
Reducing leakage by Gate = 14nm Standard Gate -> 16nm Wimpy Gate (two options; wimpy costs masks)

22FDX/22FDX+ Multimedia core
Reducing leakage by Vt&Vbb = RVT -> RVT w/ RBB -> HVT -> HVT w/ RBB (two options with boosters; no added process steps)
Reducing leakage by Gate = 20nm -> 24nm -> 28nm -> 32nm -> 36nm options (five options, no added masks)
Siginifigantly lower power thus longer battery life when doing the ole Youtube/Netflix/Disney+/Amazon Prime/etc binge shuffle.

External factor: GlobalFoundries has increased demand for 45nm SOI and 22nm SOI platforms. In turn, GlobalFoundries supply for 14nm platform is reduced as it isn't in demand nor is it contributing to income.
Internal factor: AMD has desire to re-enter low-cost via ultra-low-power CPU(AMD64/RV64)/GPU(Post-GCN)/SoC(Multimedia, Crypto-Security, RF, Integrated I/O, Vision, Semi-custom IP).
Combined factor: AMD/GF going forward partnering up on 22FDX onwards.
Proof of risk-aversion:
1. AMD removing exclusivity of 14nm/12nm node. (GloFo revenue for 14nm/12nm is less secure)
2. AMD getting lower price of 14nm/12nm node. (GloFo revenue contributes less to income)



With a return of micro-server targets because of ULP architecture and lower costing node.


I have seen the above but with fanless 4x 22FDX CPU/GPU/APUs with an eight slot cluster motherboard.
With Topaz target allowing for equivalent of: 6 CUs * 4 dies * 8 cards => 192 CUs at ULP arch/clocks.

Lower cost of entry => Higher volume of sales to consumers
Reduced cost of fabrication => Higher margin of profit to AMD

Pre-2015:
HP and LP are on the same node with low-cost option having reduced margin.

Node	HiPerf - HiCost	LowPower - LowCost
28nm	Viable	Viable
20nm	Viable	Non-Viable
14nm	Viable	Non-Viable
7nm	Viable	Non-Viable

2016-2018 Post-2019:
Splitting HiPerf and LowPower to separate nodes

High Performance Architecture + Increasing ASP, Increased Margin	Ultra Low Power Architecture + Recession Tolerant, Increased Margin
HP = 14nm/12nm	ULP = 22FDX || 2022+ :: 12FDX RiskProd
HP = 6nm/5nm || 2022	ULP = 12FDX || ~4 year cadence = 2026+ :: 6FDX RiskProd
HP = 4nm/3nm || 2024	ULP = 6FDX || ~4 year cadence = 2030+

cIOD/sIOD 12nm is easily replaced by bin harvested cIOD/sIOD 6nm. There is also second source on 16nm/12nm at TSMC becoming first-source going forward.
Pollock 3015e = July 2020 ~4 year cadence => July 2024
28nm end of manufacturing = July 2021 ~2 year substitute => July 2023
A6-9220C/A4-9120C = January 2019 ~4 year cadence => January 2023
22FDX insert is January 2023 through July 2024.

2017/2018/2019 = A55/A65/E1 [A65/E1/Helios was originally meant for 2015-20nm]
2021+ = A510, no A610/E2 OoO SMT core yet.
2025+ = A530 Hypothetical, A630/E3 Hypothetical
^-- Target of ULP AMD64/RV64

AMD ~2.7 million units for HP cores, whereas previous low-cost cores shipped ~6 million units per quarter.

Without the rush to get rid of margins to the next node by the prior low-cost core:
1.
"Lynx Family" 40nm Bobcat(2011)/28nm Bobcat(2012)
"Panthera Family" 28nm Jaguar(2013)/20nm Leopard(2015) || "Puma Family" Puma = GF28 and Catamount = GF20
"Leopardus Family" 16nm/14nm Oncilla/Margay(2017)
2.
22FDX ULP1 (2022+)
12FDX ULP2 (2026+)
6FDX ULP3 (2030+)


HP core - Increased costs from FinFET, increased TDP from increased transistors leading to higher Pstat/Pdyn/Ptotal, higher cost of entry: low customer count.
ULP core - Lower costs from FDSOI, decreased TDP from reduced transistor count lower Pstat/Pdyn/Ptotal, lower cost of entry: high customer count.

 

[itsmydamnation]

Nosta your redacted doesn't pass the simplest of ocams razor tests.

easy answer was already provided in this thread, Milan IOD

also still waiting for tunnel borer


Profanity isn't allowed in the tech forums.

AT Mod Usandthem

 

[nicalandia]

I personally don't think Monet is worth it but AMD thinks otherwise?

Maybe you should be AMD CEO? You might be smarter than Dr. Su right?

 

[jpiniero]

Maybe you should be AMD CEO?

That'd be cool. Imagine all the stock you'd get!

 

[scineram]

Maybe you should be AMD CEO? You might be smarter than Dr. Su right?

We’re not, actually. Since she canned it.

 

[NTMBK]

We’re not, actually. Since she canned it.

Wait, is Monet not happening now? I lost track of the rumours.

 

[Joe NYC]

Wait, is Monet not happening now? I lost track of the rumours.

Maybe it is:

https://twitter.com/i/web/status/1476225266800373768

https://twitter.com/i/web/status/1476229628071223301

 

[NostaSeronx]

Wait, is Monet not happening now? I lost track of the rumours.

1. Monet Never Existed
2. If it did, it would directly compete with Mendocino;

https://twitter.com/i/web/status/1420851817248923654

CxTxMFG = "Picasso/Winston" APUs by the way, nothing new which is Zen+ and Vega.

Meanwhile...


TSMC 5nm - TSMC 6nm - TSMC 7nm - GFS 22nm => No new designs on GFS 14nm/12nm or on 28nm.

It is not particularly difficult on the LP core side to get something smaller with similar IPC:
Jaguar 16nm = 2+2 => 4mm2 < 5.5 mm2
Jaguar 14nm = 1.79+1.79 => 3.58mm2 < 5.5 mm2
The area efficiency for CMP is not sufficient enough for a smaller processor. Hence, CMT's execution cores.

Lower price has higher volumes, more worthwhile GlobalFoundries push.
14LPP/12LP/12LP+ peaked at 6x volume, with 45nm SOI entering the volume has dropped down to 2x.
22FDX/22FDX+ should be at 10x volume, with no volume loss.
~20% and declining revenue node generation w/ negative contribution to net income at GF -> ~40% and increasing revenue node generation w/ positive contribution to net income at GF

A new product at GlobalFoundries 14nm/12nm FinFET would get hit by the Malta pivot. Requiring more space for more mature nodes as Fab10 customers move to Fab8.

While 22FDX between 2021 to 2023 is expected to go the farthest at 20x volume. This is not including potential 10x at Malta and 10x at Singapore as well.

New 12LP+ SoC with backported IP => December 2022 - January 2023 timeframe => No HVM capacity for 12LP+.
New 22FDX(+/++) SoC with new ULP IP => January 2023 and later timeframe => Increased capacity w/ larger market base.

"Declining personal computing market SAM due to continued and expected transition of certain customer designs to lower nodes. GF optimizing and prioritizing wafer capacity for higher profits and more durable demand from other end markets"
PC to IoT/Embedded/Home/Cost-essential PC => Compute isn't key, power efficiency is and at low cost.

 

[xilli_fiberbit]

Tesla will use Ryzen 1st gen (Zen+) CPUs until 2025. Good deal for AMD.

Part of the new slots will be used to produce IO chips for Epyc, part of the new slots will be used to produce the CPUs for Tesla.

https://twitter.com/i/web/status/1476492330010128389

 

[NTMBK]

"Declining personal computing market SAM due to continued and expected transition of certain customer designs to lower nodes. GF optimizing and prioritizing wafer capacity for higher profits and more durable demand from other end markets"

This quote right here is why you have things backwards. GloFo made plans to change wafer allocation because they expected AMD to reduce their orders! If AMD actually increases their orders of FinFET wafers, GloFo will keep producing them.

 

[Abwx]

Seems that Tesla now use an AMD set up for all models :

Tesla goes AMD Ryzen for the 2022 Model Y and Model 3 infotainment systems in America, gets rid of 12V lead-acid batteries

North America is finally getting its new 2022 Tesla Model Y and Model 3 batches with upgraded media system processors and new 12V batteries. Tesla is now equipping all new cars with AMD Ryzen chips instead of the older, slower Intel Atoms, and swapping the 12V lead-acid batteries that were prone...

www.notebookcheck.net

 

[NostaSeronx]

This quote right here is why you have things backwards. GloFo made plans to change wafer allocation because they expected AMD to reduce their orders! If AMD actually increases their orders of FinFET wafers, GloFo will keep producing them.

1. GlobalFoundries will not continue FinFET production when it has a negative contribution towards revenue.
2. AMD is expected to reduce their orders as they move to TSMC.
3. The quote is correctly interpreted.

AMD doesn't have capacity at GlobalFoundries for FinFETs. GlobalFoundries wants to turn around Malta. The whole 14nm/12nm FinFET only at Malta strategy didn't work and they never produced positive net income. So, GF said get with the times and move to FDSOI or lose capacity.

CMOS is down
FinFET is down
FDSOI/RFSOI is up
SiGe/SiPh is up

FinFET with a lower price => less money for GlobalFoundries => GlobalFoundries is now obligated to kill off FinFETs since margin has been reduced. #WeDontDoLowSlashNegativeMarginEither

This already stuck -> "22FDX Meteor" at GlobalFoundries Dresden, majority of the left over capacity ~550K is going to be 22FDX.
This will strike -> "12FDX Meteor" at GlobalFoundries Malta, majority of the existing capacity will be turned over to 45nm SOI with the 12FDX 2019-2021 mentions indicating 2022-2023 insert of 12FDX.
Singapore => ~450K increase w/ it also being 22FDX(Replacing Chengdu Phase2)

+550K at Dresden <== SOITEC Bernin II, ~850K+continual ramp FDSOI capacity
+450K at Singapore <== SOITEC Pasir Ris, ~1000K+continual ramp FDSOI capacity
+100K at Malta for Fab10 customers, with more exisiting capacity shifting to fit in 22FDX/12FDX. <== GlobalWafers MEMC & Prior commitment to SOITEC Pasir Ris. ~ramping FDSOI capacity

Just 2022 models => V180F+Navi23
While 2023+ models => R3000(6nm Mendocino) or V3000(6nm Rembrandt)+Navi23S(6nm Navi23)
Infotainment systems aren't automotive.

There is little no reason for AMD to continue using Worst Merit FinFETs on the market. When they don't have an exclusivity agreement. With the reduction of prices, there is less money to no money(>100% likely even before) for a continuous transistor booster. I highly doubt AMD will push out a GloFo 12nm auto-embedded(requires AGx Standard Libs) when they can get more capacity and lower prices at TSMC's 12nm(also, requires AGx Standard Libs).

Malta Example:

FinFETs are sold at fire/clearance sale prices till completely removed.

 

[NTMBK]

I give up. All I'll say is that I'll be astounded if AMD produce a single product on 22FDX or 12FDX.

 

[Roland00Address]

Maybe I do not understand the rest of the thread (not literally everyone, just a lot of people here =P )

But 14nm derivatives such as 12nm is the new 28nm. 14nm is still good hardware even if we are also doing much better stuff simultaneously. AMD will easily be able to use it for the next 3 years either with cheap SOC for things like chromebooks, embedded, etc but also things such as IO dies and so on which you might as well mix and match.

So my opinion summarized is this. What price did AMD get? For those 12-14nm commitments, for price and secure availability (where you do not get bumped for another customer) is what truly matters.

Likewise AMD should also be either on the leading foundry from TSMC or the 2nd best depending on the precise details. But using GF 12nm when it makes sense absolutely makes sense if AMD can get a good price and availability.

 

[NostaSeronx]

It's not very astounding though.

Again, LP and HP cores before 2015 were on the same node.
1. Both cores saw increased wafer prices; 28nm -> 20nm -> 14nm -> 7nm
2. Both cores saw increased design costs; Plus increased combined HP-o DTCO for HP and LP-o DTCO for LP
3. Both cores entered into the node at the leader.
4. Opposite customers;
- Low power SoCs have higher volume requirements but provide lower ASPs
- High performance SoCs have lower volume requirements but provide higher ASPs
- - Leader inserts at lowest capacity and early PDK.

Going forward I'll name them HPC cores and ULP cores for example.
1. Both cores are on different nodes; HPC on FinFETs and ULP on FDSOI.

2. Both cores enter a particular node at different times;
- HPC inserts at Leader (0 years to 3 years from volume introduction)
- ULP inserts at Follower (3 years to 6 years from volume introduction)
- - Follower inserts at stable high capacity and mature PDK.
;;14LPP only had 10.5T/9T at first, to only get 7.5T later on
;;22FDX only had 12T/8T at first, to get 9T-ULV, 6.75T-AG1, etc.
;; ;; of that 12T-UHP was proven not to provide good enough PPA, which 8T-UHP within 8T-ULP actually out performed.
;; ;; ;; 8T-UHP = 0.2Vmin to 1.8Vmax voltage range.

3. FDSOI-line maintains high control of Speed or Leakage.
- FinFET loses control of Speed or Leakage.

4. FDSOI has a trend of keeping up with compute efficiency at lower costs
- 22FDX is ~85% the cost of 14nm/12nm FinFETs with same transistor count
- - 12FDX is ~70% the cost of 7nm/6nm FinFETs with same transistor count

4. The ULP SoC line is definitely not targeting same transistor counts.
- Less transistors equals bigger cost savings to AMD going forward.

Hence:
Client-side = Bristol -> Raven2 -> Dali -> Mendocino (128-bit DDR line) :: Increased Performance
Embedded-side = R-series -> R1000-series -> R3000-series
Increased wafer costs but smaller or similar die at higher ASP.

Client-side = Stoney -> Stoney Refreshes -> 22FDX ULP SoC (64-bit DDR line) :: Increased Power-efficiency/cost
Embedded-side = G-series -> G-series -> G-series
Decreased wafer costs but smaller die

 

[burninatortech4]

I'd be completely shocked if there isn't a Zen 3 successor to Dali. It doesn't make any sense for there not to be. Especially when they're buying this volume of wafers. Call it Monet or don't - doesn't matter - I still think this chip is real. Time will tell.

 

[NostaSeronx]

Maybe I do not understand the rest of the thread (not literally everyone, just a lot of people here =P )

But 14nm derivatives such as 12nm is the new 28nm. 14nm is still good hardware even if we are also doing much better stuff simultaneously. AMD will easily be able to use it for the next 3 years either with cheap SOC for things like chromebooks, embedded, etc but also things such as IO dies and so on which you might as well mix and match.

So my opinion summarized is this. What price did AMD get? For those 12-14nm commitments, for price and secure availability (where you do not get bumped for another customer) is what truly matters.

Likewise AMD should also be either on the leading foundry from TSMC or the 2nd best depending on the precise details. But using GF 12nm when it makes sense absolutely makes sense if AMD can get a good price and availability.

14nm is not the new 28nm. It remains more expensive than 28nm/22nm products. Especially so, if there is no shrink.

AMD doesn't exactly have 12nm or 14nm commitments, since in the May 2021 WSA removed them.
"The A&R Seventh Amendment also removes all prior exclusivity commitments and provides the Company with full flexibility to contract with any wafer foundry with respect to all products manufactured at any technology node."
The price for 14nm/12nm stayed the same for May 2021, and decreased for December 2021.

Price is probably equal to TSMC since exclusivity was removed. Secure availability will probably only last up to 2023.

We can look at other WSAs:

Document

Single Company, Multisource:

2.6 GF shall have the exclusive right to decide at which of Fab 1 or Fab 7 (or both) a new Product will be taped-out, qualified, and manufactured. For new Products, GF shall provide Customer [***] months notice prior to tape-out of the Fab(s) that such Product will be manufactured in, [***].

For Products in production, GF shall provide Customer [***] months notice before moving any Product from Fab 1 or Fab 7 to the other GF fab, and Customer shall consent to such move except on the grounds of a good faith reasonable commercial justification for refusing such move, in which case, GF shall remain obligated to supply such Product from the original GF fab. Qualification wafers, for such Product to be moved between fabs, shall not count towards any Annual Reservation quantities [***].

If production will be in both Fabs, Customer and GF will make commercially reasonable efforts to qualify the Product at both Fabs.

With this the removal of exclusivity makes sense. Since, GloFo has no intent to increase FinFET capacity given May 2021 timeframe. With October 2021, GF announcing that they are switching away from FinFETs to SOI at Malta; with it finishing up in 2023. Where the availability will most likely shrink to zero.

14LPP/12LP/12LP+ = Only at Malta M1, w/ no expansion, only a reduction planned
16nm/12nm = TSMC Nanjing(Now), TSMC Taiwan(Now), TSMC Europe(Future)

12LP+ => 0.5V target for Std. Cells. FoM: highest leakage of the 16nm-12nm nodes.
N12e => 0.4V target for Std. Cells. FoM: lowest leakage of the 16nm-12nm nodes.

I expect it will be the opposite of Jaguar -> Puma. Zen/Zen+ is extra leaky on 14nm/12nm at GlobalFoundries. While the variants at TSMC being able to hit TDP targets promised in concept.

I'm also sure the crowd that got A8-7680/A6-7480 will get a kick out of Diffused in China/Made in China.

Foundation IP Selector

Find the best Memory Compiler, Non-Volatile Memory (NVM), and Logic IP solutions for your SoC design needs, by simply selecting your desired foundry process node

www.synopsys.com

14nm/16nm gate lengths only

16nm Standard Cell Libraries optimized for Low Power, High Performance & High Density - Dolphin Technology

Dolphin Technology provides SoC designers with Standard Cell Libraries optimized for low power, high performance and high density across a broad range of process technologies.

www.dolphin-ic.com

16nm, 18nm, 20nm and 24nm gate lengths

12nm Standard Cell Libraries optimized for Low Power, High Performance & High Density - Dolphin Technology

Dolphin Technology provides SoC designers with Standard Cell Libraries optimized for low power, high performance and high density across a broad range of process technologies.

www.dolphin-ic.com

adds 12nm gate length => Lowest Vt will use this for highest speed and the increased gate lengths above can control leakage better and reduce size/variation.

TSMC = 65% of revenue is below and not at 22nm.
GloFo = 75% of revenue is above and not at 14nm.

Keeping it minimal costs, no redesign whatsoever => Samsung has S1 Line and S2 Line

 

[naukkis]

I'd be completely shocked if there isn't a Zen 3 successor to Dali. It doesn't make any sense for there not to be. Especially when they're buying this volume of wafers. Call it Monet or don't - doesn't matter - I still think this chip is real. Time will tell.

Zen3 is 7nm design. Porting it to 12nm would cost as much as any new design on 12nm and take one cpu design team away from doing 5nm/3nm designs. So it still is extremely unlikely that we see Zen3 on 12nm node, it doesn't make any sense financially at those small production numbers. AMD does have 14/12nm cpu designs - Zen/Zen+ so every product coming from 14/12nm production lines will most likely be based on them. Zen3 @12nm backport won't be any upgrade, it mostly could be sidestep like Rocketlake.

 

[jpiniero]

Zen3 is 7nm design. Porting it to 12nm would cost as much as any new design on 12nm and take one cpu design team away from doing 5nm/3nm designs. So it still is extremely unlikely that we see Zen3 on 12nm node, it doesn't make any sense financially at those small production numbers. AMD does have 14/12nm cpu designs - Zen/Zen+ so every product coming from 14/12nm production lines will most likely be based on them. Zen3 @12nm backport won't be any upgrade, it mostly could be sidestep like Rocketlake.

If they want to stay in the cheap laptop market they have to do something. I doubt TSMC 7 nm is gong to be cheap enough any time soon so the alternative to doing something like Monet is getting out.

 

[naukkis]

If they want to stay in the cheap laptop market they have to do something. I doubt TSMC 7 nm is gong to be cheap enough any time soon so the alternative to doing something like Monet is getting out.

Zen3 backport to 12nm would cos about same as 100 million Monets produced on 7nm. So if it is about money they won't even think Zen3 on 12nm. For cheap 12nm products AMD will use what they had on 14nm - Zen+.

 

[andermans]

If they want to stay in the cheap laptop market they have to do something. I doubt TSMC 7 nm is gong to be cheap enough any time soon so the alternative to doing something like Monet is getting out.

I'd think Mendocino is in that space and is based on TSMC 7(?) nm.

 

[jpiniero]

I'd think Mendocino is in that space and is based on TSMC 7(?) nm.

I have a feeling that if there's an cut product coming it's more of an Alder (Raptor?) Lake-M competitor than the space Dali/Picasso is in.

 

[Zepp]

Is there any financial benefit for AMD to simply port Dali to a smaller node rather than designing a zen2/zen3 based replacement for low power mobile devices going forward?