Search the Community
Showing results for tags 'tsmc'.
-
Quotes The sluggish sales of Apple's MacBook have already been predicted. Apple's Mac PC business revenue in the first quarter was $7.7 billion. This is a 30% decrease compared to the same period last year. Source: https://wccftech.com/m2-chip-production-stopped/
-
Summary Taiwan is experiencing a water drought across the whole country, enforcing water restrictions in central Taiwan as reservoir supplies are dangerously low. Taiwan experienced a significant drop in rainwater in 2020, which was amplified by the fact that there were no typhoons that made landfall in Taiwan. TSMC believes the restrictions will not affect their production, as they will increase the usage of water from tanker trucks. Quotes Taiwan has issued their first red alert on water supply in over six years, as water reservoir supplies are running dangerously low. Water will be cut to non-industrial areas across central Taiwan starting April 6, 2021. Taiwan Semiconductor Manufacturing Company (TSMC) as well as Micron Technology Incorporated, have chip-manufacturing operations in central Taiwan, particularly Taichung. Taiwan's economic minister Wang Mei-hua stated that they will cut water supply to areas of Taichung, Miaoli, and Changhua by 15%. Tap water will also be suspended in these areas for two days per week, but will not apply to industries. Wang Yi-feng, the deputy director-general of Taiwan's Water Resources Agency, stated, There are other semiconductor fabrication companies in Taiwan, such as AUO, Winbond, Innolux, and GlobalWafers, whom of which have factories in Taiwan. If the drought becomes bad enough, they can affect their manufacturing capabilities, worsening the global chip shortage. Water level capacity at Taiwan's reservoirs are the following: Shimen reservoir - 49.13% Te-Chi reservoir - 10.19% Tsengwen reservoir -15.22% The Shimen reservoir in northern Taiwan is their largest water catchment reservoir. Gokul Hariharan, managing director and co-head of Asia technology, media and telecom research at JPMorgan, stated that the water shortage around the second quarter of every year is quite common and coincides with semiconductor manufacturers ramping up chip production. Taiwan's seasonal rain and replenishment of reservoirs happen towards the end of Q2 (May-June) or in the summer of Q3. Hariharan believes that production won't stop, Taiwan believes that the tech industry has enough water to last until May 2021. My thoughts Wow!!! Can it get any worse??? I feel bad for all of those that are wanting to upgrade their gaming rigs or just need new components in general. Hopefully, the water supply crisis in Taiwan improves and does not affect semiconductor manufacturing, or else the chip shortage will get very bad in an already strained market. Their predictions on the water supply crisis not affecting chip manufacturing during a time semiconductor manufacturing is at 100% will "likely" not stop production, but let's see how this will pan out. Sources https://www.techspot.com/news/89062-taiwan-introduces-further-water-restrictions-amid-drought-potentially.html https://www.bloomberg.com/news/articles/2021-03-24/taiwan-raises-red-alert-over-water-cuts-supply-for-chipmakers https://www.cnbc.com/2021/03/08/taiwan-water-shortage-president-says-to-brace-for-a-shortage.html
-
I hope someone else hasn't done this yet. Summary Reports of chip shortages impacting all manner of industries abound, it’s not just graphics cards that are hard to come by. Now comes news that megafoundry TSMC is prioritising production of chips for cars over other applications. What hope is there for us poor gamers ever getting our hands on a new GPU? According to CNBC, TSMC plans to prioritise new production capacity in favour of car makers. The move comes as pressure reportedly mounts at government level for TSMC to prioritise chips intended for cars. The global auto industry is currently suffering from significant production delays due to the shortage of chips, leading Germany’s Economy Minister Peter Altmaier writing to his Taiwanese counterpart requesting that TSMC, a key supplier of semiconductors for cars, focus on production for cars above all. Similar requests have reportedly been sent from the US, the European Union and Japan. Quotes My thoughts Well, I am personally not gonna be making any GPU purchases in the foreseeable future but don't panic, please don't. This also seems to not affect the current production capacity but future one's, so....... I'm not too worried. Sources https://www.pcgamer.com/tsmc-to-prioritise-production-of-chips-for-cars-not-your-next-graphics-card/ https://www.cnbc.com/2021/01/25/taiwan-ministry-says-tsmc-will-prioritize-auto-chips-if-possible.html
- 45 replies
-
- tsmc
- graphics cards
-
(and 3 more)
Tagged with:
-
In an interview for Time, TSMC Chairman Mark Liu, suggested that the global chip shortage is now being exacerbated by dealers stockpiling chips “who-knows-where in the supply chain”. Quotes My thoughts This again is bad news for anyone who wants or needs to get their hands on any new (or by trickle down, used) piece of technology, name it video card, car or any in between. And this extra variable of why the shortage exists is even harder to solve because it deepens speculative behavior. In other markets commodities have laws intended to stop speculative hoarding and other similar behavior. Lets see if regulations need to be taken to try to stop or disincentivize this practice. Sources Global chip shortage: Supply chain hoarding now a factor, says TSMC chief Inside the Taiwan Firm That Makes the World’s Tech Run
-
[IMPORTANT]: DO NOT INVOLVE POLITICS AND FOCUS ONLY ON THE TECHNOLOGY SIDE. Summary China is one of the largest consumer's of energy. In the effort of reducing it's emissions and reducing price surges related to energy carriers, China has decided to temporarily shutdown power to several manufacturing centers. China plans on limiting electricity on a scheduled basis going into the future. As of 2019, 64-65% of China relied on coal-fired electrical power plants to produce the electricity. Last week, President Xi Jinping vowed at the UN General Assembly to not build a new single new coal-fired electrical power plant. Quotes As a result, this will reduce CO2 emissions in the area and commodities such as coal, oil, and natural gas to help energy producers maintain low prices and help stabilize their business. Since a large chunk of electronics are produced in China, this will put extra strain on the already strained electronics and semiconductor industries. Other companies will have different ways of handling this issue, Some manufacturers will not be shutting down due to how their business operates, such as the manufacturing of semiconductors. As a result, TSMC and UMC semiconductor fabrication plants will continue to operate as normal. My thoughts This is going to be an interesting future now with the power limits China is placing throughout these manufacturing regions. With the current semiconductor shortage, perhaps this can allow companies such as TSMC the time to produce the silicon ahead of the rest of the packaging materials they sit on. Although, I don't know if those companies are affected by the shut down too, which would result in a very messy situation moving forward with supply chains. Sources Tom's Hardware BBC News ABC News
-
Summary US trade sanctions against selling AI chips to Chinese customers are set to go into full effect in March of 2023. This has left manufacturers of these chips, such as Nvidia, scrambling to sell as many of these high-margin products as they can in China before that deadline. To help accommodate them, TSMC has started offering batches of "super hot" runs, where they claim turnover time has been cut from an average of six months to around three. Quotes My thoughts I wonder if this would require TSMC to delay other orders, like AMD's next-gen products Sources https://www.tomshardware.com/news/nvidia-tsmc-rush-orders-before-china-sanctions
-
Summary New semiconductor fabrication plants in the USA (Intel, TSMC, Samsung) will have a hard time finding skilled employees to staff their plants when they are operational. The lack of a 'pipeline' and the competition and attractiveness of SWE doesn't make the job any easier for the industry. Quotes My thoughts I found this story, attributing the skill shortage to a lack of 'exposure' et al. interesting, after Linus mentioned that ALL the major semiconductor makers declined his request for a tour of their 'cool' fabrication plants and the potential for free PR & marketing. The other thing ('pipeline') is that it needed a crisis on several fronts, and an act of congress (loads of government money), to patch over some gaping holes in America's comparative competitiveness, and the semiconductors talent 'pipeline' (Vocational Education and Training). Compare this to mainland Europe and how close Universities and their leading industry companies, big and small (Mittelstand), collaborate to maintain their basic science lead and skills availability. And the entrenchment and history of VET in the European economy. Sources Protocol: New chip fabricating plants will need tens of thousands of skilled workers who don’t currently exist. Training them means persuading students to look away from jobs at big tech companies.
-
Summary While we're already seeing the fruits of Intel's collaboration with TSMC, it seems that Intel is willing to go much farther than just TSMC as a source of chips for its product portfolio. Intel CEO Pat Gelsinger recently took a trip to South Korea's capital of Seoul to visit Samsung. According to the Korea Herald, Gelsinger met several key Samsung executives; like Samsung Electronics Vice Chairman Lee Jae-yong. Quotes My thoughts I think this is a good move by Intel because it allows them to not be solely dependent on their own fabs. As we saw in the past, this proved to be problematic when they were having difficulty with their 10nm process (now called Intel 7). Tom's Hardware mentions that Gelsinger was also in Taiwan in April trying to secure more of TSMC's sub-7nm production. TSMC is now the fab being used for NVIDIA RTX 4000-series video cards (RTX 3000-series was Samsung) as TPU mentions, meaning Samsung might have an opening for clients to fill manufacturing capacity; that's where Intel comes in here. Intel Meteor Lake and Arrow Lake (2023-2024) feature a TSMC N3 (3nm) GPU, as detailed in one of their most recent roadmaps. Intel Lunar Lake and beyond might be where Samsung steps in for 2024+. Samsung could even have won some orders before Lunar Lake. Sources https://www.techpowerup.com/295406/after-tsmc-intel-may-be-edging-closer-to-samsung-for-collaboration https://www.techspot.com/news/94772-intel-ceo-pat-gelsinger-visited-samsung-discuss-potential.html https://www.tomshardware.com/news/intel-pat-gelsinger-meets-samsung-lee-jae-yong
-
Summary The TSMC and Samsung foundries are planning on price increases. For TSMC they are expecting the price increases in 2023, for Samsung the price increases sometime in the second half of 2022. Price increases between the two will vary from 8-30% according to reports. Quotes My thoughts Well this is a real bummer, right when we started to see GPU prices equalizing it appears that prices for all gaming products are still to go up (directly from the fab though). This will prove to be an issue for people trying to build new PCs, because this doesn't even take into account the behavior we've been seeing regarding prices in recent history (scalping, inflation, raw materials and lack of availability). Sources https://www.engadget.com/samsung-raising-chipmaking-prices-131536580.html https://www.tomshardware.com/news/tsmc-warns-clients-of-up-to-9-price-hike-in-2023 https://www.techpowerup.com/294753/tsmc-said-to-be-planning-price-increases-in-2023 https://www.techpowerup.com/294837/samsung-foundry-considering-up-to-20-percent-price-hikes
- 21 replies
-
Summary If a new rumor is to be believed, TSMC is set to formally announce its 1.4nm-class technology in June. The initiation of R&D on TSMC's 1.4nm chip fabrication process / pathfinding of TSMC's 1.4nm node is almost underway. Quotes My thoughts It's crazy to think of a 1.4nm process coming into play while we are currently on 7nm for AMD, 8nm for NVIDIA, and 10nm for Intel (Intel 7). Next-Gen NVIDIA will be 5nm, next-Gen AMD GPUs will be on a mix of 5nm and 6nm, next-Gen AMD CPUs will be on 5nm and next-Gen Intel (Raptor Lake) will still be on 10nm (Intel 7). While the first 2nm chips should come around 2026, they are saying expect the 1.4nm products to come around 2028. While that appears like quite a while, six years from now to be exact, it seems like only yesterday that the Sandy Bridge i7-2600k and Bulldozer FX-8150 was launching on 32nm, and the Kepler GTX 680 / Tahiti HD 7970 was launching on 28nm. The big question is what happens after 1.4nm? Do we start transitioning to different materials other than Silicon (Graphene, Carbon Nanotubes, Indium gallium nitride, Gallium-arsenide) or will progress stagnate (and we see the true end to Moore's Law)? Sources https://www.tomshardware.com/news/tsmc-to-initiate-14nm-process-technology-rd http://www.businesskorea.co.kr/news/articleView.html?idxno=92813
- 19 replies
-
Sorry for the long post, but there was so much to cover from the article and i felt this needed to be covered. I'm really rooting for the low cost, very low power technologies such as fd soi. Sure, the bleeding edge finfet stuff is sexy and all, but it's expensive. Cost for designing and manufacturing chips are going through the roof, and stuff like iot needs to be cheap and reliable in order to be adopted. Source :http://www.anandtech.com/show/11337/samsung-and-tsmc-roadmaps-12-nm-8-nm-and-6-nm-added/4
-
TSMC has announced the delivery of the complete version of its 5nm design. This means that companies like Apple and Qualcomm can begin development of their next generation of SoCs. The Apple A-Series has always been industry leading and it will be great to see what power efficiency gains can he had by going from 7nm to 5nm. I'm sure Qualcomm can show a great hustle to catch up to the A-Series as well. In my view it's unlikely that we will see much of any focus on performance improvements and rather 5nm will be focused entirely on bringing down the power cost and heat output of the chips. SoCs like the A12 Bionic are already far more powerful than any average consumer could ever hope to utilize, so it only makes sense that Apple and others will continue to refine their "high efficiency cores" to bring their single threaded performance up while reducing or keeping power consumption the same. It does appear though that 5nm will have to be validated in some way and TSMC has committed to working closely with their partners when designing new silicon: We will not see any 5nm silicon on any 2019 phones however, 5nm is slated for 2020 as of right now, which is around the time when 5G will be relevant. Source: https://www.cultofmac.com/618004/apple-can-start-designing-its-5nm-a-series-chips/#more-618004
-
Summary Fresh rumors out of South Korea has it that Samsung has snagged an opportunity to become the Sole Manufacturer and Provider of the new Qualcomm Snapdragon 875 SoC. The reason for this according to an Industry Insider is simple: Money. Samsung has simply offered lower pricing for SoCs manufactured under its 5 nm EUV process than TSMC did. The deal is claimed to be worth a cool $840 Mil. Quotes My thoughts So now that Samsung produces the Snapdragon 875 and the Ampere GPUs it'll mean that TSMC will have more fab space available for folks like AMD and Apple, so I hope for AMD that they won't experience such delays again like it happened to them with Renoir and TSMC not being able to produce enough Chips due to High Demand. And with Nvidia now acquiring ARM for $40 Billion it'll reshape the Industry of SoCs and chips based on ARM big TIme. Sources TechPowerUP
- 19 replies
-
So......right off the bat Tom divulges that two of his 'Nvidia insiders' tentatively verified that all production of "GA102 and higher" (engineering samples aside) leading up to the launch of consumer Ampere cards this fall will on Samsung's 8nm process. This falls on the heels of Intel's troubles (and Nvidia earlier this year) securing capacity at scale with TSMC as Apple, AMD, and Qualcomm have snatched up the vast majority of it. This seemingly includes the RTX 3080 (GA102), 3070 (GA103), 3060 (GA104), 3050 (GA106), and 3040 (GA107: Maybe we'll see a "Super" refresh on TSMC next year. At least, one can hope.
- 31 replies
-
- 7nm vs. 8nm?
- nvidia
-
(and 2 more)
Tagged with:
-
SOURCE: https://wccftech.com/amd-7nm-vega-gpu-zen-2-cpu-mass-production-tsmc/ TSMC has just confirmed that they've begun Massproduction of their 7nm lithography, confirming that AMD will be contracting them to produce their Vega+ GPUs as part of their Radeon Instinct and Radeon Pro lineups. These GPUs are expected to hit market during H2, 2018. AMD is also expected to use TSMC for their upcoming Zen 2, 7nm, processors. TSMC's production capacity of their 7nm lithography is expected to increase three-fold by the end for the year, this is what WCCFTech says about it: "This means that TSMC will be able to deliver a mass total of 1.1 million units by the next year, a three-fold increase over the current year. The AMD 7nm OEM orders are expected to being production in Q4 2018 and TSMC is expected 7nm to account for up to 20% of their Q4 2018 revenue. As for other nodes, TSMC will be beginning mass production on the 7nm+ using EUV technology next year, which will further reduce power consumption by 10%, compared to 7nm."
-
Freaking awesome news for mobile tech! It would appear that this massive investment is a move to secure TSMC as Apples primary provider for their A series SoCs. Historically Apple has used a combination of Samsung and TSMC, but Apple has been burned in the past using Samsung chips with the A9 SoC and has since moved to only using TSMC. It would be great to see Samsung increase their R&D budget to compete with TSMC for Apples business. Another good thing about this is manufacturers like Qualcomm are going to get the trickle down benefits from Apples desire to increase performance in their phones. So really everyone wins! Source: https://www.cultofmac.com/556985/apple-invests-25-billion-in-future-a-series-chips/
-
Today, GlobalFoundries announced that it is immediately halting all 7nm development and will not move forward with any 7nm node production and will instead focus on specialized processes. Additionally, AMD has announced that it's moving all production to TSMC in the wake of this announcement by GF. I think it's interesting to see them make such a rapid turn in such a short time. Whether this will mean better processors for other products though is still yet to be seen (and probably won't for a couple years). I hope AMD can continue pushing forward with their 7nm process at TSMC without much issue. Article: https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development Edit: Re-read the article and it looks like AMD designed their 7nm process for TSMC anyway so this shouldn't have much, if any, impact on their 7nm process (hopefully)
- 78 replies
-
- globalfoundries
- tsmc
-
(and 2 more)
Tagged with:
-
https://www.anandtech.com/show/12677/tsmc-kicks-off-volume-production-of-7nm-chips News is actually from yesterday, which surprised me that it hadn't made it into the forum. This is definitely one of those "big deal" announcements. With all of the trouble that the foundries have been dealing with in the last 10 years of nodes, this is going to be a fairly sizable move for the entire industry. By TSMC's own numbers, the move from 16FF+ node (Nvidia's Pascal is on this node) to 7FF is over a 70% area reduction. (Same number of transistors fits in 30% of the previous space. That's a massive improvement.) There are also a few other important areas this effects. We've had those PS5 rumors running around, and both the Xbox & PS4 are manufactured by TSMC using AMD designs. If 7nm node is in high volume production, then a PS5 could actually launch this year. The other really interesting tidbit is the part I bolded about "server CPU". Is this going to be Qualcomm CPUs for ARM servers, or are we going to see AMD moving Zen2 production to TSMC? This level of area shrink & performance uplift means the next 2 years of products should see a large jump in performance. Last important possibility that comes to mind is that Nvidia will be skipping their 12nm node (customized for them for Volta) and moving to 7nm for their next GPUs, which would launch later in the year.
-
During TSMC's tech symposium, TSMC unveiled their new Wafer on Wafer silicon stacking technology which, if implemented, will allow for the stacking of two silicon wafers on top of each other to connect chips as opposed to having the multiple dies sitting side by side and using an interposer. According to the Overclock3d article, the ideal situation to use this would be on wafers with chip yields greater than 90% and to decrease thermal risks, the tech is better suited for low-power parts. https://overclock3d.net/news/misc_hardware/tsmc_reveals_wafer-on-wafer_chip_stacking_technology_-_wow/1 It might be interesting to see if AMD &/or nVidia might use this to interconnect their GPUs directly to the memory. Basically stick the low power/heat silicon on the bottom and have a direct connection to the high power/heat silicon that's on the top... Might also be fairly useful for something like image processors and/or upcoming phones. I doubt that we'll see full GPU dies directly connected to CPU dies anytime soon due to heat/power constraints (although that could be interesting).
-
Here I am, back with crypto news, because of course. https://cointelegraph.com/news/china-mining-co-bitmain-shows-higher-2017-profits-than-us-gpu-giant-nvidia-report-finds Bitmain, the Chinese company behind the famous Antminer ASICs for Bitcoin, BCH, Litecoin and others, and an important client of TSMC during the last and current year, appears to have taken in profits on par or higher than Nvidia itself. I'm not up to date in the who-is-who of news outlets, but the main source is CNBC here, for what that's worth. https://www.cnbc.com/2018/02/23/secretive-chinese-bitcoin-mining-company-may-have-made-as-much-money-as-nvidia-last-year.html There may be a calling for some monopolic activity on the side of Bitmain, as they control a very high market share A lot of us, myself included, were hoping for something that stopped the insane shortage that we have in terms of consumer and enterprise silicon because of this demand for mining hardware. A lot of you want this to just die, whereas I was looking for non-minable currency to become the new big thing. Unfortunately it seems like we're gonna have to wait more. You damned miners, how dare you buy my silicon! I'm not entirely comfortable with the place the space is in right now. Especially because of environmental concerns. I hope that in the years to come, preferably sooner rather than later, minable cryptos step down and non-minable ones like Proof-Of-Stake Cardano take over the top in terms of market cap.
- 42 replies
-
- cryptocurrency
- tsmc
-
(and 4 more)
Tagged with:
-
TSMC has recently break ground for a brand new FAB located in Southern Taiwan Science Park. It will be FAB #18. Fab 18 Basic Information: Lot Size: 42 hectares Total Floor Area: 950,000 square meters Total Cleanroom Area: Over 160,000 square meters (approx. 25 standard soccer fields) https://www.techpowerup.com/240966/tsmc-breaks-ground-on-5nm-fab-18-in-taiwan https://hexus.net/business/news/components/114695-tsmc-started-construction-5nm-fab-18-taiwan/
-
Much as I have found AdoredTV's channel to be grossly biased and misrepresentative, I came across a rare gem today, and it got the gears in my head spinning. In summation, Adored thinks Intel is positioning itself mostly to damage Nvidia in mobile and reap more profits from the sector along the way, and he believes Nvidia has focused on datacenter so much in recent years BECAUSE it saw the long-term trend of the industry with integration and already had terrible relationships with Intel, Apple, and OEMs. I'm in agreement with all of this, but I thought a little bit farther outside this microcosm, and I'd like some feedback and discussion. To start, Intel EMIB (Other Parts Incoming Post-Discussion) EMIB allows Intel to connect any harmonious group of IC dies, including CPUs, GPUs, DSPs, FPGAs, DRAM, and Non-Volatile Memory in a manner far more compact, more scalable, and cheaper manner than using interposers. We've already seen how Intel intends to leverage this by directly linking AMD Polaris and Vega GPU dies to HBM2 alongside Kaby Lake and Coffee Lake CPUs in a custom compact AIO setup for Apple and likely other laptop OEMs as highlighted here: https://www.anandtech.com/show/12003/intel-to-create-new-8th-generation-cpus-with-amd-radeon-graphics-with-hbm2-using-emib As Adored points out, this serves to undercut Nvidia's dGPU pricing and cut them out of the mid-high and down mobile range; and he does point out Intel could easily play AMD and Nvidia against each other to compete on pricing, but I disagree with his stance that Nvidia would refuse to take the loss on selling more cheaply to Intel. It's better to lose a little than lose it all, tactically and strategically speaking. I think Nvidia will compete for scraps even as it pivots more and more to Datacenter and AI. However, this DOES also represent a serious problem for AMD. While it is still significantly cash-strapped and lacks the mindshare both in PC and mobile to puts its graphics tech back on the map in the minds of consumers, Intel still has the IPC lead in CPUs as well as the overall cheaper manufacturing purely given its merit of operational size. If AMD can't get Raven Ridge and Zen or Polaris and Vega into high-end mobile, this represents a better-than-nothing deal. Adored also thinks AMD is shooting themselves in the foot somewhat since they could turn around and do exactly this with interposers, especially since Zen seems to have the power and clocks advantage in mobile. However, the fact they entered into this deal suggests mindshare IS the blocker here. With Intel's endorsement, perhaps this partnership will only last a year or two before AMD can deliver an all-AMD AIO solution with mass appeal outside enthusiast circles. However, there is still the problem of price. Interposers are NOT cheap. They are strictly bigger than the sum total of areas for each die in the package. They are also thicker than EMIB-based solutions, where the interconnect is literally embedded in the substrate of the daughter board or mainboard. And while some remain hopeful EMIB can be opened up as prior art via a patent dispute lawsuit, my learned opinion is not optimistic for that. What do you think? Will Nvidia compete for space in the new Intel AIO package era, or simply give away midrange mobile? Will thin and light dominate so much that interposer solutions are unappealing long-term? Will Intel be able to guard its embedded bridge designs from competitors? What are your thoughts on Intel leveraging EMIB tech in its custom foundry to possibly get a piece of the ultra mobile pie (phones and phablets getting even thinner and lower power)? Do you see Intel's "CPU/APU" strategy in the next 3-5 years manifesting to include non-volatile storage (Optane/NAND) and HBM/HMC caches in packages which themselves drop into consumer boards? What interesting applications do you see for EMIB outside the "APU" sphere? Part 2: Intel/AMD Graphics We know Intel has access to all the Maxwell and previous IP from Nvidia until the end of time thanks to its patent deal, for which all of the payments have been made. We also know Raja Koduri was poached by Intel and will be heading up a discrete graphics division for the company. We also know Raja was responsible for Vega and Navi pretty much start to finish. And we know AMD's graphics division actually has not been doing so well in recent years (see Adored's videos concerning the GPU wars and AMD's semi-custom strategy for details, taking careful note of AMD's bare 1 million USD profit on the graphics division in 2014/15). With this information added to the semi-custom graphics chip deal being supplied for this EMIB package and unknown specifics regarding Intel's access to the driver stack, are we potentially looking at the early stages of a deal between AMD and Intel to hand over or spin off Radeon back into its own company for the two to share and help resource? Is there an additional IP deal between Intel and AMD which is aiding them in their work to create their own dGPUs? With Raja having left after taking a sabbatical and an internet beating by fans for Vega's disastrous delays and launch and incomplete driver support to this day for tiled rendering and other features, could there have been an agreement between AMD and Intel to begin a transition over with investor politics wanting to cut its losses/dead weight with graphics? If a handover were going to happen, who better to manage it than the man who headed the division under AMD? Perhaps the timing is coincidental, but in this industry, that's rarely the case. It seems strange to me that this all strategically aligns for Intel and also seems to alleviate a black eye that AMD has been suffering since it bought ATI in 2005 and has only just finished fiscally paying for. What's more is AMD's driver team is known to be quite small. Intel has an army of programmers at its disposal and is very active in its open source Linux driver development. What are your thoughts on this? Extremely convenient coincidence? Perhaps there's another angle being played where AMD retains full control of Radeon moving forward, but it seems like it would be unpopular with the Board of Directors at best.
-
https://www.semiwiki.com/forum/content/5533-asml-imec-euv-progress.html http://semiengineering.com/euv-cost-killer-or-cost-savior/ http://semiengineering.com/gaps-remain-for-euv-masks/ http://semimd.com/blog/2015/10/02/spie-photomask-panel-money-is-an-issue/ The semiconductor industry unfortunately has a very bitter pill sitting in front of it which may have to be swallowed. In spite of the progress made in the field of Extreme Ultra-Violet Lithography, the next-gen semiconductor production technology is nowhere near ready for deployment, and deadlines are approaching at speeds that don't seem viable. And what's more, this technology was only expected to last for 3-4 node shrinks at most, and that was for the original phase-in goal of 10nm. For reference, the reason current lithographic technologies are at their limit is quite simple: light resolution. The current etching technology uses 193nm light. How do we get "14nm" circuitry you ask? Mirrors, and special smoke and lenses and masks and multipatterning. In simpler terms? Money. Lots of money paying lots of engineers to support legacy hardware to the bitter end and buy time. This is why the semiconductor foundry industry has been reduced to just 4 major players with fat pocketbooks: Global Foundries, Taiwan Semiconductor Manufacturing Company (TSMC), Intel Corporation, and Samsung Semiconductor. These four huge companies are responsible for the vast majority of semiconductor production around the world. Recently even the titan IBM has pulled out of this business due to rising costs and lack of profit margins. So why does EUV solve so many problems? That light is at a much more...native resolution you could say, of 13.5nm. It 's much easier to shape light in small degrees and proportions compared to what we force 193nm light to do these days. The problems lay in generation for long periods of sustained time and having "capture" hardware that can withstand Ultra-Violet lasers for sustained periods of time. But, all the money in the world cannot defy the laws of physics; and unfortunately that means numerous attempts to find viable solutions to persistent aforementioned problems have been expensive at best and financial black holes at worst. In 2012 alone Intel invested 4.1 billion USD into ASML to help accelerate research for 450mm wafer tooling and EUVL. Fab 42 to this day (Intel's only 450mm plant, meant to build 14nm products) remains dormant, and EUV isn't on track to meet anyone's deadlines. http://www.bloomberg.com/news/articles/2012-07-09/intel-agrees-to-buy-10-stake-in-asml-for-about-2-1-billion Samsung donated another ~1.4 billion to this in the same year, and the money has only continued to pile up and burn. https://www.asml.com/asml/show.do?ctx=5869&rid=46974 For reference, we already have Intel and others using multiple patterning at 14nm, and keeping production costs down and yields up has been a struggle since 22-20nm. 10nm was never going to be easy after we saw the stall in Intel at 14nm, but while the dogs are starting to go down, the world has come around ready to start kicking. In terms of cost reduction, there's just no way to do it even now. Sure the number of production steps may decrease moderately from what we have currently(remember 14nm requires about 1/3 the steps that 7nm does for 193nm Liquid Immersion Patterning), but the equipment even now has lifetimes far too short and fidelity just not up to snuff. For reference, a standard 193nm power source (already at 150-180W) used for something like Samsung's or Intel's 14nm processes lasts about 3000 hours, and the replacement time is about 4 hours since the instruments are much less heavy and complexly interconnected. And the mirrors can tolerate many more imperfections and do not wear out nearly as quickly. Further, the mask infrastructure and supply lines (think of these as inverted maps which reflect and refract light onto the wafer in order to make the correct etchings, and bear in mind the sort of material must stand up to what is essentially an ultraviolet laser for extended periods of time) show no signs of improvement. Where production times fall, facilities have to be built modularly at every step to keep the lines flowing when machine parts inevitably fail. Integrating what many consider a necessity to progress is proving to be a logistical nightmare. Beyond that, the tools and components will be much more expensive and die faster. The final nail? Neither ASML nor IMEC have projections which Intel believes will have the tech ready before 2019. At that point an additional 1-1.5 year stall to retool facilities will have officially killed Moore's Law without an intervening miracle. The choices beyond that point are all unpleasant in their own ways. AMD already employed HDL at 28nm for both GPU and CPU silicon. Whether we see other foundries follow suit will be borne out soon enough. Late last year IBM demoe'd a 7nm wafer built using 4-deep multiple patterning and experimental EUV on Silicon-Germanium using both FinFET and FDSOI design principles, supposedly an experimental Power 9 or maybe even Power 10 chip design, emphasizing progress may not be possible in upcoming years unless a massive breakthrough is made. The industry is on-edge, and other options for innovation are now being considered which no one likes. Intel has already said clock speeds will have to necessarily fall for it to increase power efficiency and stay in-line with ARM's improvements, but what will this entail? Most don't disclose too much about their Silicon processes, so perhaps one or more of these options has already been used. Fully-Depleted Silicon Techniques (Samsung is employing these in their FinFET-based 14nm LPP process which GloFo will use for AMD's production) Silicon-On-Insulator (FDSOI support from many companies) 3D FinFET (Samsung, Intel) FinFET Silicon on Insulator (IBM) Quantum Well Transistors (Experimental MIT, UC Berkeley, Intel) Tunnel FET (Experimental Japan and South Korea) Spintronics (Experimental various) New Materials (Silicon-Germanium at SUNY, Albany with IBM and GloFo, Indium-Gallium-Arsenide at Samsung, III-V Materials at various locations, carbon nanotubes at IBM and MIT, and graphene at MIT and UC Berkeley) Die Stacking (I really don't see this as viable outside of memory technologies for thermal reasons primarily, but maybe I'm wrong) Clearly innovation will have to continue once we reach 10nm, but it seems like an industry stall is inevitable due to lithography technology failing to have a significant breakthrough in the last five years. How each foundry addresses this stall will be pivotal to the survival of each. In this age of aiming to reduce power consumption, my money is on FDSOI first given the infrastructure is already in place. Beyond that I just don't know enough to say with any significant degree of confidence.