Best guess is multiple chip (4 64 bit wide) Custom very high density stacked RAM for a total of 256 bits wide on substrate called Chip on Wafer on Substrate like in this picture:I'm hoping the rumor of Xbox next having HMC is true
It would be hilarity
jeff_rigby said:Both PS4 and Xbox 720 are using the same or similar technology made possible by AMD, IBM and others in the consortium. This is a LEAP in technology that has been in the planning since 2008 by everyone!I think it would be a better idea to split this thread for Xbox 8 and PS4. It would be easier to read the tech findings for each console in their own threads?
Microsoft-sony.com domain registration seems to imply a partnership which I'm speculating is shared R&D and setup costs for Game Console SOCs either Xbox361 & PS3 refreshes coming at the end of this year or Xbox720 & PS4 or possibly both. Differences would be outside the core SOC...memory size, multi-media support etc.
It's now pretty much confirmed that both Xbox 720 and PS4 will be using exactly the same AMD building blocks and possibly the same core silicon. Parts of the speculation is confirmed for next generation but my speculation that it may also apply to refreshes of this generation (PS3 4K chassis) has been proved wrong.This thread will become one epic thread of Jeff talking to himself, until the PS4 launches.
Which means lot of conversation.
HPC memory and graphics memory GDDR5 post for (High Performance Computing) is heading to the introduction of the technology of wide memory interface. It uses a technology solution: (Through Silicon Via TSV) through-silicon vias. However, it is said instead of DRAM stack directly on top of the CPU and the GPU, and are considering the introduction of a method to use the I / O chip and silicon interposer.
One of the plans that has emerged is a way to connect with a wide DRAM interface CPU and GPU silicon interposer technology using TSV. Called for the 3D stacking DRAM stacked directly on CPU and GPU, and how to use the interposer is 2.5D. As in the case of 3D stacking, ultra-wide interface can be used by a very large number of micro-bump pins. However, unlike 3D stacking, there is no need to use a TSV the GPU or CPU, in terms of reduced risk of manufacture.
By using ultra-wide interface, low power consumption, using the TSV solution can achieve ultra-wideband memory. Standards relatively quiet, "Wide I / O" is a wide interface for mobile DRAM, the bandwidth is 12.8GB/sec memory per chip. However, in ultra-wideband 4-8 times of Wide I / O, in a non-mobile, memory bandwidth is that you want to target are considering to 100GB/sec per chip. Development of Wide I / O These are sometimes referred to by the name Wide I / O and Ultra Wideband Wide I / O, such as computing Wide I / O.
Sonys next game station logic- on-interposer will reportedly similarly be fabbed by Global Foundries, and packaged by a collaborating OSAT (probably ASE, StatsChipPAC, SPIL or again Amkor).
Earlier in the year a panel of manufactuers (including Global Foundries), talking about 2.5/3D infrastructure, were asked if any of them would be supplying Sony for PS4.
Interesting, rumors are 300FPS (from Sony CTO) 10 eye views (5 S3D) at 30 FPS each and the target of 10X GPU performance. I guess this is where the specs for next generation come from for Sony. They want to support glassless 3-D on 4K TVs with the PS4.2010 April 22 System Software Ver3.30 Every Single PlayStation® 3 (PS3®) supports Stereoscopic 3D (S3D)
1. Game Industry struggle for seeking the new exciting experience
2. Reality is the critical trend of Gaming
3. Interactive S3D brings you to next Level
4. Do not mind wearing 3D Glasses Game is the Concentration-required Media
5. Relatively Easy to add S3D functionality to 3D Game
5 Major Characteristics to explain Graphics Evolution
3D is the key word of Modern Game 3D Graphics
S3D Graphics World’s First Real-time 3D Graphics Game Console END-to-END 3D Graphics Improve the quality of Real-time 3D Graphics “Reality”is the key trend of the Home console games
FPS (First Person Shooter)
Free World / Sports / Driving
Games Become Real-World Simulator thanks to the improvement of Quality in Graphics S3D Adds new dimension of Reality to the Game
S3D with Glass 2x Render Performance 2 eye views
Naked-eye S3D 10x Render Performance 5 2 eye views
http://eandt.theiet.org/magazine/2011/12/maasaki-tsu-interview.cfm said:Sony’s target is to get latency for a typical playing experience to below 50ms for framerates of more than 300fps. Now, 50ms is an absolute best performance level to start with – most displays actually increase it – for framerates of about 60fps ceiling. Moreover, the target is not for 1080p resolution, but reflect a drive towards 8kx4k.
This is designed for Handheld SoCs and currently expensive. AMD has 20nm process now, some parts @ 20nm may make it into the PS4. Heat issues in 3D stacking are less with more efficient smaller node sizes. A wait to use more 20nm process might be attractive enough to delay a launch to early 2014. AMD and GloFlo thinking outside the box and mixing process optimized building blocks in both 3D/2.5D stacking of discrete silicon and on the same building block silicon (14nm transistors on 20nm plane).14nm finfet technology will level the playing field for the fabless community by giving them a process as good as anything available in the IDM community, says Mojy Chian, senior vice president for design enablement at Globalfoundries.
Globalfoundries intends to have 14nm finfet in volme manufacturing in 2014, the same timescale as Intel has for introducing 14nm finfet manufacturing.
In fact, GF's 14nm process may use smaller transistors than Intel's, said Chian because "Intel's terminology doesn't typically correlate with the terminology used by the foundry industry. For instance Intel's 22nm in terms of the back-end metallisation is similar to the foundry industry's 28nm. The design rules and pitch for Intel's 22nm are very similar to those for foundries' 28nm processes."
On the Intel 22nm process the drawn gate length is actually 26nm.
GF's 14nm process delivers a number of benefits over 20nm. "One benefit is that it reduces development time because people can easily migrate from 20nm to 14nm," said Chian, "customers can use the same GDSII and swap the transistors." GF calls it 'fin-friendly migration.'
Fin-friendly migration benefits from the characteristic of GF's 14nm process that it retains the 20nm interconnect while moving to 14nm transistors.
Other benefits of GF's 14nm are that it will deliver 20-55% higher performance than 20nm depending on the operating voltage, or that it can deliver 40-60% better battery life with the same performance.
The GF 14nm technology uses rectangular-shaped fins. That could change when it moves into volue manufacturing when the exigencies of real-world production require compromises, but GF's current expectation is that the rectangular shape will survive into manufacturing. Intel has had to go to a triangular-shaped fin which reduces the performance and the power-saving benfits of finfets.
At the front of the GF 14nm finfet node will be mobile SOCs. GF has been working with ARM since 2009 to optimise its processes for ARM-based SOCs. 14nm will be the first process at which Intel will put mobile SOCs to the front of the node.
By moving to 14nm finfet while keeping 20nm interconnect, GF has brought forward the intorduction of its 14nm process by one year.
Next year GF starts production on 20nm. GF's 20nm SRAM, and various customer test chips on 20nm, are yielding "very well", said Chian.
The above meets the need of a Game console "the performance should be stretched at the launch time" and multiple Forges can produce it. (GloFlo & IBM are currently 28nm gate first so designs for them can't work with TSMC. At 20nm AMD is moving to gate-last and that design can be produced by TSMC.)http://www.guru3d.com/news_printer/amd_20132014_and_2015_gpu_codenames.html said:According to an article, AMD will continue the trend in 2014 with "Volcanic Islands". This GPU series will go head-to-head against NVIDIA's Maxwell and Intel's second-gen Xeon Phi architecture. Volcanic Islands will be the first GPU family from AMD to be manufactured at 20nm Gate-Last process, giving the company freedom to choose between the Common Platform Alliance (IBM, GlobalFoundries and Samsung) as well as long-term partner TSMC. This will also be the year when we'll see full system integration between the CPU and GPU. The APU (for consumers and servers) will extend to the discrete graphics card and treat it as an integral part, and the other way around.
- Owned the register file cell qualification analysis in the 22nm node
- Owned the design of a 3.2GHz one read one write custom register file array in the 22nm node.
I don't disagree and value your input (many of your older BY3D posts showed great insight). I'm confused now as it appears that GLoFlo is ready to start 14/20nm next year. Is next year Dec 2012?To have any kind of volume for a launch between September to November, chip production has to start at let 3 months earlier, probably 6.
Even if 20nm were ready for production next summer, the volumes will not be there for a mass market consumer device.
We maybe looking at a quick shrink in 2014 for the consoles however.
http://eda360insider.wordpress.com/2011/12/14/3d-week-driven-by-economics-its-now-one-minute-to-3d/ said:According to the data gleaned from presentations by Samsung, Toshiba, AMD, and others, 3D IC assembly gives you the equivalent performance boost of 2 IC generations (assuming Dennard scaling wasn’t dead). Garrou then quoted AMD’s CTO Byran Black, who spoke at the Global Interposer Technology 2011 Workshop last month. AMD has been working on 3D IC assembly for more than five years [2008] but has intentionally not been talking about it. AMD’s 22nm Southbridge chips will probably be the last ones to be “impacted by scaling” said Black. AMD’s future belongs to partitioning of functions among chips that are process-optimized for the function (CPU, Cache, DRAM, GPU, analog, SSD) and then assembled as 3D or 2.5D stacks.
- Owned the register file cell qualification analysis in the 22nm node
- Owned the design of a 3.2GHz one read one write custom register file array in the 22nm node.
http://news.softpedia.com/news/AMD-Tests-32nm-Starts-Working-on-22nm-98400.shtml said:the latest news on the web announces that the company has also started working on the 22 nanometer technology, the next step after 32nm. This is a long shot move, as the technology will not be ready before 2012, while products manufactured under the process should surface in the first half of 2013.
According to the news, AMD is working on the 22nm process with IBM, which may prove to be beneficial for both companies.
http://www.brightsideofnews.com/news/2012/3/14/ibm2c-globalfoundries2c-samsung-to-go-finfet2c-3d-transistor-with-14nm-process.aspx said:The alliance will offer FinFET to all of its customers at the 14nm process node, which paired with Fully Depleted Silicon-On-Insulator (FD-SOI) is pairing incredible transistor density with lower power consumption. The problem with the current partially-depleted SOI (PD-SOI) technology is that the pressure needed for SOI insulation to take place is decreasing yields due to pressure on the already strained silicon. When you pressure the strained silicon transistors, they tend to break. This is also the main reason why Common Platform Manufacturing Alliance, manufacturing arm of Common Platform Technology Alliance decided to go SOI less and go bulk with the 28nm and 20nm processes.
Furthermore, the FinFET will be combined with chip stacking technology, which in a very high-level overview - will lead to a very different semiconductor products for the world of tomorrow. We will cover the todaý's event in greater detail as the week unravels.
Don't forget that consoles are doing more and more these days. When the 360 and PS3 launched they just played games for the most part. Now both (especially the 360) have had endless features added to the OS, almost to the point where it can't handle it. Next generation Sony and Microsoft are going to need to set aside a decent amount of ram for the OS running in the background. Hell look at the Vita. While it's not as powerful as the PS3, it does have more RAM because of what the OS demands. 2gb doesn't seem like enough for both a large OS and next gen games. I expect 4gb of ram, at the least.So you're in the camp that would rather see more slower ram in next gen consoles because it's a higher number, rather than less but faster ram? You may just get your wish.
This thread will make a lot more sense when the next generation officially begins in a few weeks and the thread title won't get changed.
This thread will make a lot more sense when the next generation officially begins in a few weeks and the thread title won't get changed.
Last year AMD officially became an ARM licensee, although the deal wasn't publicized at the time. Fast forward to June 2012 and we saw the first fruits of that deal: AMD announced it would integrate ARM's Cortex A5 core into its 2013 APUs to enable TrustZone support.
Today comes a much bigger announcement: AMD will be building Opteron processors based on a 64-bit ARM architecture. There are no product announcements today, but the 64-bit ARM Opterons will go into production in 2014.
The only other detail we know is that these ARM based Opterons will embed SeaMicro's Freedom Fabric, presumably on-die.
I suggest reading the article Part the Clouds: AMD Fusion APUs Ideal for Cloud Clients and this continuation of the article. Lots there that applies to us and it's an easy read. After reading think about Zlib having hardware support in AMD APUs and the W3C supporting WebCL and Zlib compression as the standards to be used by web browsers. Think about AMD also including a ARM A5 for DRM and more in their APUs.Man, I would love some of this in English.
Departure from SOI
At the Morgan Stanley Technology, Media and Telecom Conference Thomas Seifert last week gave a bit more detail on the matter:
"We said that at the 28nm node we are going to be on bulk silicon across all products, not only graphics but also CPUs. And We have made no statement beyond that. But for 28[nm] we will be on bulk for all products."
He also added:
"There are always tradeoff decisions. But the flexibility that we gain moving in that direction... the flexibility across foundry partners, across design tools outhweigh that by far, the benefits of SOI."
EDIT: So this 20nm wafer (IBM-GloFlo) had to be made before January 2012 (test). The 20nm " APUs coming in the 2013 timeframe" will not be Kabini which is at 28nm. So there will be a 28nm APU and 20nm APU out in 2013 but there is no mention from AMD of the 20nm APU name or specs, could it be the "Confidential simi-custom win" read game console?http://www.brightsideofnews.com/news/2011/4/27/amd-fusion-tapeouts-unveiled-28nm-in-2q-20112c-20nm-in-2012-and-14nm-in-2014.aspx said:For the end, we leave you with an image of a 20nm wafer which was shown to us in January of this year, clearly showing that a factory alliance led by IBM and GlobalFoundries has a "horse to run" in the semiconductor manufacturing arena. This process will be the base for AMD's Fusion APUs coming in 2013 timeframe:
22nm SOI Process Integration starting at GloFlo the same time Microsoft and Sony started next generation (2010) with Katherina Babich moving from IBM to GloFlo. In this time period Sony announced they will be skipping the 32nm node and another Linkedin IBM post has research on 22nm cell register arrays.SMTS
GLOBALFOUNDRIES
Privately Held; 10,001+ employees; Semiconductors industry
2010 – Present (2 years)
22nm SOI FEOL Process Integration
Senior Engineer Scientist
IBM T.J. Watson Research Center
Public Company; 10,001+ employees; IBM; Information Technology and Services industry
2007 – 2009 (2 years)
Process Integration for 22nm FINFET and High-K Metal Gate Devices:
• Led and enabled raised source and drain (RSD) module development work to support critical 22nm and beyond device programs
• Developed robust low temperature selective Si Epi growth process for 22nm multi-gate nonplanar devices FINFET and Tri-gate devices.
FD-SOI bests FinFETs for mobile multimedia SOCs? ST says yes.
Tuesday, November 29th, 2011
In a recent and excellent article in ASN by Thomas Skotnicki, Director of the Advanced Devices Program at STMicro, he explains in a very clear and accessible way why FD-SOI with ultra-thin Body & Box (UTBB) is a better solution for mobile, multimedia SOCs than FinFETs — starting at the 28nm node and running clearly through 8nm. It is based on the paper he presented at the 2011 IEEE SOI Conference.
Consider the improvements in performance that ARM’s seeing on an M0 core on 20nm FD-SOI vs. 28nm bulk: 40% better at 1V, 56% at 0.9V, 81% at 0.8, and an amazing 125% better performance at 0.7V.
As SOI Consortium Director Horacio Mendez pointed out in ASN this summer, you typically expect to get about a 25% improvement in performance moving to the next node. But ARM’s showing that if you move to the next node and move to FD-SOI, you get really phenomenal results, especially at the lower supply voltages.
This is a big find. I suggest others take the time to read through the patent. It's all about affecting devices based on user(s) gaze (eye tracking).SCEA R&D's apparent love affair with eye tracking also continues:
http://www.freepatentsonline.com/20120300061.pdf
That's several eye tracking patents they've filed over the last 18 months or so, ranging from the tracking itself through application in games and non-game software.
Almost reminds me of when a bunch of 'rear-touch' patents popped up in the run up to Vita's announcement...but again, who knows.
Pennar/Samara (2014 Jaguar APU) ported from 20nm TSMC database to Global Foundries. It has 2 128 bit LPDDR3 memory interfaces and a New GNB. It fits in the same product lineup as Kabini. My guess is that Pennar is the near future AMD APU design being used by the PS4.AMD Southern Islands: Radeon HD 7950 (TAHITI PRO), Radeon HD 7970 (TAHITI XT), Radeon HD 7990 (NEW ZEALAND), IBIZA, COZUMEL, KAUAI.
AMD London series: WIMBLEDON, PITCAIRN, HEATHROW, CHELSEA.
AMD THAMES, LOMBOK, GREAT WALL, SUMMER PALACE, CAPE VERDE.
AMD Sea Islands: Oland, Bonaire, Hainan, Curacao, Aruba.
AMD Solar Systems: Mars, Sun, Neptune, Venus.
AMD Fusion: SUMO, WRESTLER, TRINITY DEVASTATOR/SCRAPPER, RICHLAND, THEBE, KRYPTOS.
AMD Kabini, Samara, Kaveri, Pennar.