TSMC Boosts Output of 28nm Chips by Nearly 70% in the Second Quarter.

Taiwan Semiconductor Manufacturing Company said that in the second quarter of 2012 sales of wafers processed using 28nm manufacturing technology contributed to 7% of the company's revenue, which means an increase of supply by nearly 70% (if actual revenues are taken into account). Potentially, it means substantial boost of 28nm supply by TSMC, even though it is obvious that the demand still by far exceeds supply.
TSMC to Continue Boosting 28nm Output

"This year, every quarter, the major effort has been to ramp up 28nm and in doing so of course we did incur a lot of costs. As a result, the gross margin of 28nm all year this year will not be up to the corporate average standard. [...] 28nm is progressing very well," said Morris Chang, chief executive officer of TSMC, during a recent conference call with financial analysts.

28nm process technology accounted for 7% of total wafer revenues, meeting TSMC's internal plans. 40nm accounted for 28% of total wafer revenues, and 65nm was 26%. The advanced technologies accounted for 61% of total wafer revenues.



"Our output and our yields are both above the plans that we set for ourselves and the plans that we communicated to our customers early in the year. [...] We expect to ramp up to about 68 thousand 300mm wafers per month by the end of the year," claimed Mr. Chang.

68 thousand 300mm wafers processed using 28nm manufacturing technology is a very high number.Back in Q1 2012 the world's largest contract maker of semiconductors mainly produced chips using 28LP (SiON) process technology, whereas companies like Nvidia Corp. blamed it for inability to supply processors manufactured at 28HP node that utilized high-k metal gate (HKMG) technology. TSMC projected to boost output of 28HP chips to 50% of all 28nm output by the end of the year. Unfortunately, the company did not reveal the split between 28nm SiON and 28nm HKMG process technologies planned today.

"First quarter of 2013 on, we will fully meet the 28nm demand. It is also then that we expect that the 28nm gross margin will catch up with the corporate average," added Mr. Chang.
TSMC Expects to Further Increase Revenues

“Due to continuing strong demand for our 28nm technology, we expect to double the shipments of 28nm in the third quarter. This increase in 28nm business will account for more than 80% of revenue growth in the third quarter,” said Lora Ho, SVP and Chief Financial Officer of TSMC.



TSMC recently announced consolidated revenue of NT$128.06 billion, net income of NT$41.81 billion, and diluted earnings per share of NT$1.61 (US$0.27 per ADR unit) for the second quarter ended June 30, 2012. Year-over-year, second quarter revenue increased 15.9% while both net income and diluted EPS increased 16.3%. Compared to first quarter of 2012, second quarter of 2012 results represent a 21.4% increase in revenue, and a 24.9% increase in both net income and diluted EPS. Gross margin for the quarter was 48.6%, operating margin was 36.5%, and net margin was 32.7%.

For the third quarter 2012 TSMC expects revenue to be between NT$136 billion and NT$138 billion; gross profit margin is expected to be between 46% and 48%; operating profit margin is expected to be between 34% and 36%.

Microsoft Admits: Own-Brand Surface Tablets May Impact Relationships with Partners.

Microsoft Corp. has admitted in a filing with the Securities and Exchange commission (SEC) that its forthcoming Surface media and productivity tablets will compete against third-party slate-type products that use Windows 8 operating system, which may lower the commitment of the company's partners to the platform.

"Our Surface devices will compete with products made by our OEM partners, which may affect their commitment to our platform," a statement by Microsoft reads.

Partners of Microsoft Corp. believe that considering the company's heritage and orientation, it will not be able to make its own-brand Surface tablets a strong success since it lacks retail presence comparable to Apple's and will compete not on its own field. To make the matter worse, it will compete against its hardware partners, which is hardly good for business in general. According to Microsoft partners among solution providers and the channel resellers, Microsoft will not be able to convince business customers to buy Surface tablets directly from them hoping that they [clients] will customize the tablets later on themselves. Meanwhile, eliminating partners by selling Surface only in Microsoft's own retail and select online stores, means that the slates will not be tailored for particular deployments in general.



Bill Gates, a co-founder of Microsoft and the former chief executive officer of the company, said in an interview that Surface would not rival products from other makers of hardware and therefore would not impact value added resellers as well as distributors.

"I actually believe you can have the best of both worlds. You can have a rich eco-system of manufacturers and you can have a few signature devices that show off, wow, what is the difference between a tablet and a PC," said Bill Gates.

One of the reasons why Microsoft intends to release Surface is to set up a benchmark for Windows RT/8-based slates, something that hardware makers will have to look at and something that will be clearly competitive against products like Apple iPad. Many hardware makers will now have to reconsider their designs in order to make their devices on par with "reference" tablets from Microsoft. From this point of view, Microsoft's move with Surface is completely justified as it will ensure that Windows 8/RT-powered tablets will be rather good. On the other hand, some partners may decide to utilize other operating systems in order not to compete against Microsoft directly.

AMD Reportedly Delays Release of Dual-Chip Flagship Graphics Card Again.

Advanced Micro Devices will delay the release of its dual-chip flagship graphics card, the Radeon HD 7990, to late August. The reasons for the delay are unclear, but what is important is that add-in-card partners of AMD do not seem to have plans to release their own-developed Radeon HD 7970 X2 ahead of the Radeon HD 7990, despite of expectations.

Although AMD has managed to release the world's first graphics processing unit (GPU) made using 28nm process technology, the Radeon HD 7900 "Tahiti", it has evidently failed to develop a dual-chip flagship graphics card powered by two of such chips. The arch-rival of AMD, Nvidia was several months late with its first 28nm GPU, but was quick enough to launch its dual-chip graphics board for enthusiasts in late April.

According to VR-Zone web-site, AMD has delayed its Radeon HD 7990 again and will unveil it at Germany's Games Convention (GamesCon) video game event in late August. It is noteworthy that this will be one of the first AMD product launches in years that will be made in Europe. In the recent years AMD launched its significant products either in the U.S. or in the Southeast Asia.



AMD Radeon HD 7990 graphics card is expected to carry two Tahiti XT (Radeon HD 7970) graphics processing units (GPUs) and 6GB of GDDR5 memory onboard. Thanks to two AMD Tahiti XT graphics chips, the dual-chip graphics solution will have 4096 stream processors, 256 texture units, 64 render back ends and so on. The novelty will have some kind of Turbo mode that will clock the chips higher when thermal design power allows. Besides, the reference design of 7990 will have four mDP display connectors as well as two dual-link DVI connectors to support up to six displays at once, something that Nvidia's GeForce GTX 690 cannot. What is surprising is that AMD is reportedly working on a special version of the graphics card with 12GB of high-speed GDDR5 memory onboard.

The exact reasons behind the delay are unclear, but besides natural challenges with installation of two Radeon HD 7970/Tahiti XT chips with ~200W thermal design power on the same board, VR-Zone claims that AMD decided to implement a rather exotic power supply circuitry. Instead of using two 8-pin PCIe power connectors, AMD decided to utilize four 6-pin PCIe connectors. Those four 6-pin connectors lead to a multiphase power regulation, divided into three groups. The major groups consist out of Tahiti XT (not the XT2, i.e. the GHz Edition), 3GB GDDR5 memory per core (clocked at 1.5GHz) and the PLX PCIe 3.0 controller.

The Radeon HD 7990 is projected to be more affordable than Nvidia's GeForce GTX 690, which official price-tag of $999 is exceeded in many cases.

AMD did not comment on the news-story.

Asrock Pre-Announces AMP Memory by AMD.

Asrock, one of the most-admired makers of mainboards according to an X-bit labs survey, said that through a new BIOS update its flagship AM3+ mainboard will support AMD-specific memory modules with serial presence detect (SPD) containing special timings for microprocessors designed by AMD.

"Asrock's flagship board Fatal1ty 990FX professional is the only on-sale motherboard offering AMD AMP memory overclocking kit in the consumer market, stepping to the future memory overclocking now," a statement by Asrock reads.



Asrock indicated that its top-of-the-range Fatal1ty 990FX Professional mainboard which has AM3+ socket that supports both current FX "Bulldozer" chips as well as forthcoming FX "Piledriver" central processing units is now also compatible with AMD's AMP (AMD Memory Profile) SPD settings which allow overclockers and enthusiasts to easily get more speed from their memory modules by choosing exact clock-speed and latency settings [which are presented] without faithfully harming the chips.

Potentially, the improvement of supporting pg AMD's AMP-compatible memory chips should enable frequencies ahead of 2.8GHz speeds (supported by Trinity officially) from default of 2.1GHz for memory, which is a rather high frequency.

Microsoft Unveils Surface Media Tablets.

Microsoft Corp. on Monday introduced its first media tablets that will be sold under its own brand later this year. The company executives showed two Windows tablets and accessories that feature significant advances in industrial design and attention to detail. Surface is designed to seamlessly transition between consumption and creation. The Surface tablets will primarily compete against Apple's iPad slates.
Microsoft Surface is designed and engineered entirely by Microsoft employees and features extensive investment in industrial design and real user experience. The casing of Surface is created using a unique approach called VaporMg, a combination of material selection and process to mold metal and deposit particles that creates a finish akin to a luxury watch. Starting with magnesium, parts can be molded as thin as 0.65mm, thinner than the typical credit card, to create a product that is thin, light and rigid/strong.

The VaporMg approach also enables a built-in kickstand that lets one transition Surface from active use to passive consumption – watching a movie or even using the HD front- or rear-facing video cameras. The kickstand is there when needed, and disappears when not in use, with no extra weight or thickness.

Microsoft Surface also features Touch Cover keyboard that senses keystrokes as gestures, enabling to touch type significantly faster than with an on-screen keyboard. Touch Cover clicks into Surface via a built-in magnetic connector. Surface sports a full-sized USB port and a 16:9 aspect ratio of the screen, which is more advanced than 4:3 aspect ratio of Apple iPad.

Two models of Surface will be available: one running an ARM processor featuring Windows RT, and one with an Intel Core "Ivy Bridge" processor featuring Windows 8 Pro. Surface for Windows RT will release with the general availability of Windows 8, and the Windows 8 Pro model will be available about 90 days later. Both will be sold in the Microsoft Store locations in the U.S. and available through select online Microsoft Stores.
Suggested retail pricing will be announced closer to availability and is expected to be competitive with a comparable ARM tablet or Intel Ultrabook-class PC. OEMs will have cost and feature parity on Windows 8 and Windows RT.

Intel Acquires 1700 Radio Patents for $375 Million.

Intel Corp. and InterDigital  on Monday announced that they had signed a definitive agreement to sell to Intel roughly 1700 patents and patent applications for $375 million in cash. The radio technology patents will help Intel to better compete on the market of cell phones, tablets and similar devices.
"These patents will support Intel's strategic investments in the mobile segment. The addition of these patents expands our already large, strong and diverse portfolio of intellectual property," said Doug Melamed, Intel senior vice president and general counsel.
The agreement involves patents primarily related to 3G, 4G/LTE and 802.11 technologies. InterDigital is an active developer of advanced wireless technologies including WCDMA (wideband CDMA), HSDPA (high speed download packet access) and HSUPA (high speed upload packet access) 3G technologies as well as LTE (long term evolution) and LTE-Advanced 4G technologies.

"This transaction, which involves a small portion of our overall patent portfolio, marks an important milestone of InterDigital's stated strategy of expanding the monetization of its large and growing intellectual property portfolio. By executing on our business plan, which has been broadened to include patent sales, licensing partnerships and other possibilities, we see tremendous potential to expand revenue and build shareholder value," said Scott McQuilkin, senior executive vice president, strategy and finance at InterDigital.
The companies expect the transaction to be completed in third quarter 2012, subject to customary closing conditions and any required regulatory approvals.

Nvidia Readies GeForce GTX 680M for High-End Laptops.

Nvidia Corp. plans to introduce its new high-end graphics solution for notebooks at Computex 2012 in Taipei, Taiwan, early next month. The new GeForce GTX 680M will be based on the GK104 graphics processing unit that belongs to the Kepler family of solutions and will be offer about one third better performance than the model GTX 670M.
According to reports from web-sites like VideoCardz and Chiphell, the GeForce GTX 680M will feature 768 stream processors and therefore 64 texture units, reduced number of raster operating units and 256-bit memory bus. It is presently believed that the GeForce GTX 680M will utilize N13E-GTX-A2 chip, which is believed to be a variation of GK104 chip. The new mobile flagship will fully support modern high-end features like DirectX 11.1, OpenGL 4.2, OpenCL 1.2, stereoscopic-3D, 4 multi-monitor capability, PhysX, 4-way multi-GPU, PCI Express 3.0 and so on.
The reports claim that the GeForce GTX 680M is 37% faster compared to GeForce GTX 670M, which is based on Fermi architecture and which has 336 stream processors operating at 1200MHz. Given the fact that the data comes without attribution to any kind of source, it should be considered as inaccurate.
Even though both Nvidia and its arch-rival AMD in the recent cycles did not use flagship GPUs in mobile computers, Nvidia did squeeze the GF100 chip into mobile thermal envelopes and introduced GeForce GTX 480M back in mid-2010.
Nvidia did not comment on the news-story.

JEDEC Publishes Next-Generation LPDDR3 Standard for Low-Power Memory Devices.

JEDEC, a leading standard setting organization for  the microelectronics industry, on Thursday published the  final LPDDR3 low power memory standard, designed to satisfy the performance and memory density needs of the new generation of mobile devices, such as smartphones, tablets, ultra-thin notebooks and so on.  Leading memory makers - Elpida and Samsung - have already demonstrated the first LPDDR3 chips.
“To help address the dramatic rise in data-intensive apps and the resulting demands on device memory, JEDEC LPDDR3 is designed to focus on higher bandwidth requirements for device processors and graphic units. LPDDR3 represents countless hours of collaboration within the JC-42.6 subcommittee, and was developed rapidly in order to meet the mobile industry’s bandwidth requirements,” said Hung Vuong, chairman of JC-42.6 subcommittee, which standardized the LPDDR3 standard.

LPDDR3 offers a higher data rate, improved bandwidth and power efficiency, and higher memory densities over its predecessor, LPDDR2.  LPDDR3 achieves a data rate of 1600MHz (versus 1066MHz for LPDDR2) through the addition of new features, including:

• Write-Leveling and CA Training, which allow the memory controller to compensate for signal skew, ensuring that data input setup and hold timing as well as command and address input timing requirements are met while operating at the industry’s fastest input bus speeds;
• On Die Termination (ODT), an optional feature that enables a light termination to LPDDR3 data lanes to improve high-speed signaling with minimal impact on power consumption, system operation and pin count;
• Low I/O capacitance.

 As with LPDDR2, LPDDR3 supports both package-on-package [stacked] and discrete packaging types in order to meet the requirements of a wide array of mobile devices, offering designers the ability to select the options that best meet the needs of their product. LPDDR3 will preserve the power-efficient features and signaling interface of LPDDR2, allowing for fast clock stop/start, low-power self-refresh, and smart array management.
Both Elpida Memory and Samsung Electronics demonstrated their first 4Gb LPDDR3 memory components operating at 1600MHz last year. Originally, both planned to start production of LPDDR3 in late 2012, but the turmoil situation around Elpida may impact the LPDDR3-related plans. Analysts predict that demand for LPDDR3 will start to emerge in 2013.
“LPDDR3 builds on the revolutionary LPDDR2 standard, which paved the way for an entire generation of high-performance, low-power mobile devices. Now with LPDDR3, JEDEC has taken the standard to a new level, and we are pleased to offer a solution for the performance demands of a new generation of mobile products," said Mian Quddus, the chairman of the board at JEDEC.

RunCore Develops Self-Destructive Solid-State Drive.

RunCore, a supplier of solid-state drives, this week unveiled InVincible solid-state drive, which is the world's first and only SSD that features physical self destruction feature. Besides typical elimination of all data through overwriting, the InVincible SSD can completely destroy the data by physically destroying memory and other chips inside the drive by burning chips down physically.

The RunCore InVincible SSDs do not offer leading-edge performance: they can read data at up to 240MB/s speed and write data at up to 140MB/s speed. The drives use Serial ATA-II interface and can use single-level cell (SLC) or multi-level cell (MLC) NAND flash memory. But the drives can operate in rather unprecedented temperature range - between -45°C and +95°C - and support brutal ways to erase the data to keep it from obtaining by unauthorized people.

The RunCore InVincible SSDs allow to protect sensitive data from third party access in two ways with a single click. One method is an intelligent elimination of all the data by overwriting the entire disk with meaningless code. This overwrite-deletion method ensures that there is no way to potentially recover previous data stored to the device, effectively setting the SSD back to factory default. Still, there are various data recovery devices not available commercially which can, perhaps, recover data even after overwriting. A less subtle method is the physical destruction of memory chips inside SSD by applying an over-current to the NAND flash memory and thereby physically destroying them.

RunCore positions its InVincible SSDs for aerospace, military and general industrial applications, it is unclear whether the company plans to sell such SSDs to general public and whether the company will be allowed to do so.

Team Group Unleashes World's First 3GHz DDR3 Memory Modules.

Team Group, a leading maker of high-performance dynamic random access memory modules, has announced the world's first and yet the only memory sticks officially rated to run at whopping 3.0GHz. The new Xtreeme Limited Edition modules have all chances to become enthusiasts' favourite memory in the coming months.
The Team Xtreem DDR3 3000 Limited Edition 4GB memory modules are designed to operate at 3000MHz with CL11 13-13-35 latency settings on systems powered by Intel Core i-series 3000 "Ivy Bridge" central processing units and Intel Z77 platforms. The modules comply to XMP v1.3 specification, hence, should work at full speed on virtually any advanced mainboard with appropriate microprocessor. The extreme DRAM sticks utilize 8-layer printed-circuit board and bifurcated heatspreaders to ensure proper signal quality and cooling.

Unfortunately, Team Group claims that its PC3-24000 memory modules will only work in pairs, hence, it is theoretically impossible to install 16GB of such ultra high-speed memory into one PC. The modules also work only in dual-channel mode, hence, cannot be installed into PCs with quad-channel memory sub-systems.
Pricing of Team Xtreem PC3-24000/DDR3 3000MHz 8GB (2*4GB) memory kits is unknown. Keeping in mind that 16GB of DDR3 memory at 2.8GHz costs $600, do not expect 8GB of DDR3 memory at 3GHz to cost less than $300.

Samsung Starts to Mass Produce 4Gb LPDDR2 Memory Using 20nm-Class Technology.

Samsung Electronics, the world's largest maker of dynamic random access memory, on Thursday said it had begun production of 4Gb [512MB] LPDDR2 1066MHz memory components using 20nm process technology. Such memory solutions should allow to increase the amount of DRAM inside smartphones and tablets to 2GB, which will enable new classes of mobile applications and will let hardware makers to increase screen resolutions on mobile devices without compomising performance.
By stacking four 4Gb LPDDR2 components in a single LPDDR2 package, Samsung can deliver 2GB multi-layer memory solutions that boast thickness of 0.8mm. This new package is approximately 20% thinner than 2GB packages that stack four 30nm-class 4Gb LPDDR2 chips.  The new 2GB package can process data at up to 1066MHz and thanks to thinner process technology promises to provide better energy efficiency compared to previous multi-layer LPDDR2 memory solutions.

“Samsung began expanding the market for 4Gb DRAM last year with the first mass-produced 30nm-class DRAM, and now we are working on capturing most of the advanced memory market with our new 20nm-class 4Gb DRAM,” said Wanhoon Hong, executive vice president of memory sales and marketing at Samsung Electronics.
Economic benefits of the new 20nm-class 4Gb LPDDR2 are projected to help speed up the growth of the 4Gb DRAM market. Samsung now expects the newly introduced 20nm-class 4Gb LPDDR2 will rapidly replace 30nm-class 2Gb-based 1GB LPDDR2 that was in limited supply at the 0.8 mm thickness.

According to IHS iSuppli, shipments of 4Gb LPDDR2 will steadily increase, taking approximately 13% of total mobile DRAM shipments in 2012, 49% in 2013 and 63% in 2014, with 4Gb mobile DRAM monolithic becoming the mainstream chip in the Mobile DRAM market around the end of 2013.
“In the second half of this year, we expect to strongly increase the portion of 20nm-class DRAM within our overall DRAM output to make the 4Gb DRAM line-up the mainstream product in DRAM production, and therefore keeping the leadership position in the premium market and strengthening our competitive edge,” added Wanhoon Hong.

Apple iPad to Maintain Tablet Market Leadership This Year - IHS iSuppli.

After suffering a temporary dip in market share in the fourth quarter of 2011, Apple Inc.’s iOS is expected to reassert its commanding leadership of the worldwide tablet space in 2012, according to an IHS iSuppli.

Apple Likely to Sustain Leadership

After dipping to 55.1% in the fourth quarter of 2011, according to a final estimate, the Apple operating system’s share of worldwide tablet sales - a segment including both media tablets and PC-type tablets—is set to recover to 61% for the full year of 2012, about the same portion it had in 2011.
Apple’s dominating media tablet market share in the fourth quarter of 2011 had been diminished by a surge in sales of Amazon’s Kindle Fire tablet, which is based on Google’s Android operating system. This had caused Android’s share of the tablet operating system market to climb to 41.1%, up from 31.1% during the third quarter of 2011. However, as Apple reasserts its leadership, Android’s share will decline to 38.4% for the full year of 2012.

“The key to Apple’s media-tablet success has been its offering of a complete hardware-plus-content ecosystem. The combination of a good-looking device, well-designed applications, video, books and music has provided consumers with an easy-to-use product and an appealing use case. Such an ecosystem took Apple years to put together, starting with the iPod plus iTunes Music Store more than nine years ago, and it’s proving to be a challenge for the company’s competitors to replicate it,” said Rhoda Alexander, director for monitors and tablets research at IHS.
Further bolstering Apple’s commanding position in the market, supply-side sources indicate that the company will deploy a smaller, 7.85" display version of the iPad later this year, although Apple has yet to confirm this. A smaller screen does not necessarily mean a substantially lower price; rather, IHS expects Apple will place continuing emphasis on the quality of the overall tablet experience and the benefits of selecting the company’s products.

Sales of Tablets to Hit 126.6 Million Units in 2012

Sales of tablets this year - including both media tablets and PC-type tablets - will soar to 126.6 million units, up a remarkable 85% from 68.4 million units in 2011. The impressive performance of tablets this year builds on an even mightier 253% explosion last year from sales of 19.4 million units in 2010. Tablets comprise one of the strongest categories in the consumer electronics market today, with heady growth in the next few years matching the wild exuberance of the cellphone or mobile handset industry in its initial years of market-busting expansion.
Tablet sales will rise another 63% next year, on their way to 360.4 million units by 2016.

Enter the PC Tablet

While media tablets such as the iPad dominate now and throughout the forecast, new ultrabook offerings and the release of Windows 8 later this year will help drive stronger sales in 2013 and beyond of PC-type tablets, IHS predicts.
PC tablets will appeal to users wanting the flexibility of a tablet with the versatility of a traditional computer. These devices are able to manage multiple windows and applications including traditional full desktop applications, but can also convert to a slate form with touch capability. The smaller, lighter form of some of the new ultrabook offerings, touch improvements in Windows 8, and more aggressive pricing will help drive growth in this category.
Media tablets are often designated as “consumption-type” products with which users can browse the web, send email, view video, play games or interact with applications.

Within the media tablet space, however, the market is fragmenting into two segments - value products largely serving as “consumption-type” portable media players; and higher-performance units incorporating more complex applications and stronger processors. Much of the growth in the future will come from the value segment, but the performance sector will provide the stronger challenge to traditional PCs in both business and consumer markets.
Overall, the growth last year of media tablets dwarfed that of tablet PCs, and media tablet sales will continue to outperform those of tablet PCs in 2012. By next year, tablet PC growth will accelerate to nearly 160%, compared to a still-robust 60% increase for media tablets.
The PC tablet growth is a form transition within the larger notebook market and does not reflect any cannibalization of the media tablet opportunity. This is because PC tablets will still lag well behind their media tablet counterparts next year, numbering a little over 8 million units compared to more than 197 million units for media tablets.

Half of Microprocessors Sold in 2011 are Hybrid Chips with Integrated Graphics - Report.

Half of the $111 billion microprocessor market was generated by hybrid processors that included graphics cores or other kinds of processing engines, according to a report by IMS Research. Hybrid microprocessors will not only continue to gain market share, but will also let companies like Advanced Micro Devices and Intel Corp. onto new markets, such as media tablets and smartphones.
“Through the last decade the mobile and media consumption device markets have been pivotal for this hybridization trend; Apple, Broadcom, Marvell, MediaTek, Nvidia, Qualcomm, Samsung, ST Ericsson, Texas Instruments and many other processor vendors have been offering heterogeneous application-specific processors with a microprocessor core integrating a GPU to add value within extremely confined parameters of space, power and cost. Now that smartphone sales exceed computers, and tablets explode onto the market, these hybridized application specific mobile processors represent the next largest class of processor by revenue," said Tom Hackenberg, semiconductors research manager and author of the Comprehensive Processor Report.
Hybrid microprocessors - which combine x86 cores and graphics processing units on one chip -  released by AMD and Intel in 2010 and 2011 were among the most important recent product introductions. This hybridization appears to be a critical step in upping the competitive edge in computers but is also a step for these vendors into smartphones, tablets and other high performance embedded devices. Hybrid applications processor growth in smartphones and tablets are predicted with upwards of a 10% and 14% CAGR from 2011 to 2016 respectively.
“With double digit revenue growth in these markets, it’s not surprising to see major processor vendors such as AMD and Intel adopting this strategy to compete for computer market share and expending increasing research and development on embedded solutions. This hybridization is getting even more competitive with an Intel Atom-based smartphone on the market this month demonstrating that Intel is serious about entering the mobile device market by combining an x86 microprocessor, graphics licensed from Nvidia and configurable security logic as a hybrid processor triple play to capture share,” said Mr. Hackenberg.
As economic and physical barriers to shrinking geometries raise concerns of the imminent demise of Moore’s Law, semiconductor providers are turning to hybrid processors as just one of many complementary technologies for increasing system level performance and adding processor value. IMS Research has identified no less than 20 processor vendors that now provide dozens of heterogeneous processing solutions on a single chip.
Some of these converging processors have been evolving over time such as the digital signal controller, a convergence of DSPs and MCUs with the real-time processing performance of a DSP and an expanded instruction set for controller applications. In configurable processors, historically those found in FPGAs, processor vendors such as Xilinx, Altera, Microsemi and Cypress Semiconductor are actively targeting SoC ecosystems with an embedded processor core identical to and applications processor or microcontroller but enhanced with configurable logic.
“This trend is growing and spreading. Other processor vendors are now including application specific configurable logic. Intel’s Z2460 mobile processor includes a configurable security engine; Analog Devices’ BF60x DSP targets the anticipated high-growth market of embedded vision with a signal processing SoC that includes configurable logic specifically accelerating vision applications acceleration,” added Mr. Hackenberg.
Computers and media consumption devices dominate revenues in the processor market easily allowing hybrid processors to capture over half the revenues in 2011, but the trend does not stop there.
“If processor suppliers are going to continue meeting performance expectations set by Moore’s Law, clearly this is a necessary trend. Future processors may be less marketable by their top-end frequency and more by their application specific cores,” said the analyst.

Intel Boosts Security Capabilities of vPro Platform.

IT managers face a range of challenges from complex business processes to sophisticated security threats. To address these challenges, Intel Corp. has added a number of new functions and capabilities into its third-generation Core i-series "Ivy Bridge" processors with vPro technology designed specifically for business and intelligent systems.
The enhancements to the Intel Core vPro processor platform provide a more secure platform for business computing and drive the next wave of innovation in intelligent systems. New capabilities embed security at every layer, including the silicon, without compromising performance. Software innovation allows IT managers to set up and configure systems within minutes to quickly implement compelling solutions. Additionally, the enhanced graphics and secure manageability help accelerate the transition and growth in intelligent systems for the retail, industrial, and healthcare industries.

To defend against identity theft, Intel introduced Intel Identity Protection Technology with public key infrastructure (Intel IPT with PKI) into Intel Core vPro processors. The technology provides a new second layer of authentication embedded into the PC that allows websites and business networks to validate that a legitimate user is logging in from a trusted PC by using a private key stored in a PC's firmware. Intel has been working with solution providers and online Web properties such as Feitian, InfoServer, Symantec and Vasco to take advantage of Intel IPT technology to safeguard users' identity.
The latest version of the Intel Core vPro platform also features Intel OS Guard and Intel Secure Key. Intel Secure Key, with Intel AES new instructions, protects media, data and assets from loss, while Intel OS Guard detects and prevents malware. The new Intel Core vPro family of chips includes Intel Active Management Technology (AMT) to remotely manage computing issues.
Intel Core i-series "Ivy Bridge" chips with vPro technology are or will be available for desktops, laptops and ultrabooks.

Chief Exec of AMD: We Will Come In and Steal Ultrabook's Bacon.

Chief executive officer of Advanced Micro Devices believes that lower cost of the company's A-series "Trinity" accelerated processing units will help it to steal a lot of notebook market share from Intel Corp., which is pushing relatively expensive ultrabooks. The firm believes that it does not need high-end client chips in the light of the fact that that the cloud computing is emerging.
"I think we come in and steal the bacon around the whole thin-and-light movement and capture a significant portion of the opportunity there. [...] That [performance client computing] era is done. There is enough processing power on every laptop on the planet today," said Rory Read, chief executive officer of AMD, in an interview with Bloomberg Businessweek.
The head of AMD was referring to the fact that A-series Fusion "Trinity" APUs is behind Intel's Core i-series "Ivy Bridge" chips in terms of general-purpose performance, but is ahead when it comes to graphics processing performance.


What should be kept in mind is that high-performance in the cloud means high-performance server processors. Since technologies used to create high-performance chips are used both for central processing units for datacenters and for microprocessors aimed at client PCs, it is impossible to supply high-end server CPUs and sell low-performance client chips: either both types show decent performance or both do not. Nowadays AMD's highest-performance FX-series microprocessors are behind of Intel's performance-mainstream Core i7-3000-series chips across the board.
Mr. Read claims that instead of boosting general-purpose performance of client CPUs, AMD needs to integrate as many functions as possible into its microprocessors, something that makers of ARM-based system-on-chips do.

Nvidia Introduces Tesla K10, Announces Tesla K20 Accelerators [UPDATED].

UPDATE: Correcting the number of stream processors in GK110 chip.

Nvidia Corp. this week formally announced two new Tesla compute accelerators that are based on the code-named Kepler architecture. The Tesla K10 - based on the well-known GK104 chip - will become available shortly and will be aimed at those, who need maximum raw single-precision compute performance here and now. The Tesla K20, which will show up late this year, promises to be a performance monster thanks to GK110 chip with whopping 7 billion transistors.
The Nvidia Tesla "K10" compute card is based on two GK104 GPUs that deliver an aggregate peak performance of 4.58TFLOPS single precision, 0.190TFLOPS double precision and 320GB/s of m memory bandwidth. The Tesla "K10" is optimized for customers in oil and gas exploration and the defense industry. Since the K10 compute board is powered by GK104 chip, which is generally meant for graphics processing, it does not deliver really strong double precision performance and will not be a good solution for many fields of high performance computing.
The flagship compute solution based on Kepler architecture will be Tesla K20, which will be based on GK110 graphics processing unit. The latter will be a monster chip containing whopping 7.1 billion transistors, 15 streaming multiprocessors with total of 2880 stream processors and "delivering three times more double precision [performance] compared to Fermi architecture-based Tesla products", according to Nvidia. Given the fact that the highest-end Tesla 2090 provides 0.665TFLOPS of DP performance, the Tesla "K20" has potential to deliver up to enormous 2TFLOPS DP, although Nvidia claims about "over 1TFLOPS" (which is not that impressive, considering the fact that AMD Radeon HD 7970 hits 947TFLOPS DP at 925MHz).
But in addition to enormous transistor count and high performance, the GK110 will offer new capabilities that are not available on other chips:

• Dynamic Parallelism enables GPU threads to dynamically spawn new threads, allowing the GPU to adapt dynamically to the data. This simplifies parallel programming, enabling GPU acceleration of a broader set of popular algorithms, such as adaptive mesh refinement, fast multipole methods and multigrid methods.
• Hyper-Q enables multiple CPU cores to simultaneously use the CUDA architecture cores on a single Kepler GPU, which increases GPU utilization, slashing CPU idle times and advancing programmability. Hyper-Q is ideal for cluster applications that use MPI, according to Nvidia.

The GK110 GPU is expected to be incorporated into the new Titan supercomputer at the Oak Ridge National Laboratory in Tennessee and the Blue Waters system at the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign.
Nvidia Tesla K20 is planned to be available in the fourth quarter of 2012.

STEC's CellCare Technology Improves Endurance of MLA Flash Memory by 13 Times.

STEC, a leading maker of solid-state drives, said Wednesday that the company has validated its proprietary CellCare technology's ability to extend the endurance of 24nm consumer-grade multi-level cell (MLC) NAND flash memory to 40 000 program/erase cycles, which is greater than 13 times the manufacturer's specified endurance metric of 3000 program/erase cycles.
CellCare Technology is a proprietary combination of hardware and firmware working together to improve MLC flash performance and endurance throughout the life of the solid-state drive. Key endurance-extending features of this patented technology include, flash parameter tracking per die, which delivers reduced and controlled wear; flash management through the lifetime, which delivers reduced wear, controlled performance, improved performance, and lower write/read latency; and advanced error correction with digital signal processing to improve data reliability and require a minimum of re-reads.
"This announcement opens up a lot of new opportunities, as cost is king in NAND and SSDs. Being able to take consumer-grade MLC NAND from 3000 cycles to 40000 cycles presents a huge leap in cost-effective enterprise endurance - enabling much more adoption," said Alan Niebel, founder and chief executive officer for Web-Feet Research.

The new CellCare technology enables development of a new generation of STEC's enterprise-class SSDs that utilizes low-cost cMLC NAND chips, while still achieving 10 full capacity random writes to the drive each day for five years, and data retention for three months (at 40°C) with no degradation in performance. This equates to approximately 7.3PB of data being written to a 400GB SSD over the life of the drive.
After applying STEC's CellCare Technology in combination with its fourth-generation ASIC-based SSD controller, STEC engineers performed a series of JEDEC-standard tests, including a program/erase endurance and data retention stress test; a stress-test-driven qualification of integrated circuits; and a SSD requirements and endurance test.
Results of the endurance testing are outlined in the table below:

"The results of the 24nm MLC endurance testing, performed at our San Diego R&D center, are testimony that the solid flash management algorithms and digital signal processing/error correction code within our patented CellCare technology bring true enterprise-class capabilities to consumer-grade flash memory. We have now tested the use of lower-cost 24nm MLC NAND with our CellCare technology to withstand the rigorous workload environments of enterprise storage applications - specifically, to provide for 40 000 program/erase cycles," said Pablo Ziperovich, STEC's vice president of flash channel development.
CellCare Technology is embedded in STEC's full line of MLC NAND-based enterprise-class solid-state solutions, including its PCIe solid-state accelerators, ZeusIOPS SAS SSDs, and MACH16 SATA SSDs. STEC's 24nm MLC NAND flash-based SSDs are currently in development and testing, with production anticipated for later this year.

WD's HGST Begins to Use 500GB 2.5" Platters in New-Generation Consumer Hard Drives.

HGST (formerly Hitachi Global Storage Technologies), a division of Western Digital, has introduced a new generation of hard drives for consumer electronics applications. The new CinemaStar hard disk drives (HDDs) utilize latest-generation 2.5" platters with 630Gb/inch² areal density and thus provide 500GB capacity while being very - 7mm - thin, or up to 1000GB in 9.5mm models.
"The ever increasing appetite for on-demand entertainment is driving the adoption of high-capacity 2.5" hard disk drives into a growing number of CE applications. Not only is there a desire to increase the storage capacity in these devices to hold more programs, but they must deliver predictable A/V streaming performance," said Brendan Collins, vice president of product marketing at HGST.

The new family of CinemaStar 2.5" hard disk drives for such applications as high-definition multi-stream DVRs, tuner-based STBs as well as IPTV set-top boxes, DVR-enabled TVs, audio systems and video surveillance systems includes C5K1000 (9.5mm, 5400rpm), Z5K500 (7mm, 5400rpm) and Z7K500 (7mm, 7200rpm) sub-families of hard drives:

• CinemaStar C5K1000 features dual-disk 9.5mm designs with 640GB, 750GB and 1000GB capacities, 5400rpm spindle speed, 8MB cache and Serial ATA-600 interface. The manufacturer declares 1.5W typical power consumption, 5.5ms average latency, 15ms typical seek time and 124.75MB/s media transfer rate.
• CinemaStar Z5K500 features one-disk 7mm designs with 500GB, 320GB and 250GB capacities, 5400rpm spindle speed, 8MB cache and Serial ATA-300 interface. The manufacturer declares 1.5W typical power consumption, 5.5ms average latency, 15ms typical seek time and 125.5MB/s media transfer rate.
• CinemaStar Z7K500 features one-disk 7mm designs with 500GB, 320GB and 250GB capacities, 7200rpm spindle speed, 32MB cache and Serial ATA-600 interface. The manufacturer declares 1.7W typical power consumption, 4.2ms average latency, 15ms typical seek time and 174.5MB/s media transfer rate.

All CinemaStar drives feature HGST’s patented SmoothStream technology, which supports the industry-standard ATA‑7 streaming command set, and SMART command transport (SCT) that provide time-limited error recovery and thermal monitoring capabilities for a longer system life. Together, the SCT protocol and HGST SmoothStream technology minimize disruptions in stream delivery by adapting buffer management behavior and error recovery timing to match the characteristics of typical streaming applications.
"With our broad new family of 2.5" CinemaStar drives, we continue to give our customers the flexibility to select a variety of reliable and robust drives to meet their cost, performance, power or capacity requirements, while helping them to deliver smaller, more aesthetically appealing designs," added Mr. Collins.
The new 2.5" CinemaStar family is now shipping in limited quantities. CE OEM samples are now available.
"Increasingly, set-top-box, DVR, and game console manufacturers are leveraging smaller 2.5" form factor HDDs instead of traditional 3.5" drives in more compact CE product designs. With 2.5" HDDs able to offer the capacity sweet spot for set top boxes and DVRs, IDC expects 2.5" HDD shipments into these applications will grow at a worldwide 2011-2016 compound annual growth rate of 21.8%," said John Rydning, IDC's HDD research vice president.

AMD Formally Introduces Second-Generation Fusion "Trinity" Chips for Notebooks.

Advanced Micro Devices on Tuesday officially introduced its new-generation of Fusion A-series accelerated processing units (APUs) previously known as Trinity. The new chips feature new x86 code-named Piledriver cores and Radeon HD 7000-series graphics cores. AMD hopes that its new APU will help it to increase presence on the market of various notebooks.
“The latest OEM notebooks, ultrathins, all-in-ones and desktops based on the new AMD A-Series APU enable the best video and gaming experiences, highly responsive performance with AMD Turbo Core, and accelerate an ever-increasing range of productivity and multimedia applications -- in sleek, stylish designs at price points that make sense,” said Chris Cloran, corporate vice president and general manager of AMD client business unit.
The AMD A-series "Trinity" APUs feature up to four x86 cores powered by enhanced Bulldozer/Piledriver architecture, AMD Radeon HD 7000-series graphics core with DirectX 11-class graphics support, DDR3 memory controller and other improvements, such as new dynamic acceleration Turbo Core technology, improved video playback engines and so on. The new chip is made using 32nm process technology at Globalfoundries, just like its predecessor code-named Llano, but thanks to architectural improvements it is projected to be 25% - 50% faster - depending on the task - than the first-gen A-series APU.

“Our second-generation AMD A-series APU is a major step forward in every performance and power dimension, allowing users to enjoy a stunning experience without having to give up the things that matter to them most. This experience doesn’t stop at mainstream notebooks. It carries over into affordable ultrathin form factors featuring the latest in AMD Radeon graphics,” added Mr. Cloran.
The initial family of AMD A-series Fusion "Trinity" chips currently includes only five models with 17W, 25W and 35W thermal design power.

AMD claims it has a record number of design wins with companies like  Acer, Asus, HP, Lenovo, Samsung and Toshiba based on its new AMD A-series APUs, with mainstream and ultrathin notebooks as well as embedded solutions, available beginning today.

Nvidia Unveils Cloud Graphics Processing Technologies.

Nvidia Corp. on Tuesday introduced what may be one of its most important developments in the recent years: cloud technologies that allow processing of graphics data from any device. Nvidia’s Grid and VGX cloud graphics processors enable gamer-grade or professional-grade graphics capabilities on absolutely any device, including desktops, laptops, tablets or even smartphones.
Thanks to a set of virtualization technologies supported by graphics processing units (GPUs) that belong to Kepler family, Nvidia now offers Grid, a cloud graphics solution for consumers for gaming applications, and VGX, a cloud graphics solutions for professional usage in professional programs. Both technologies enable high-performance graphics processing on devices that simply cannot take advantage of the latest power-hungry GPUs.
"Kepler cloud GPU technologies shifts cloud computing into a new gear. The GPU has become indispensable. It is central to the experience of gamers. It is vital to digital artists realizing their imagination. It is essential for touch devices to deliver silky smooth and beautiful graphics. And now, the cloud GPU will deliver amazing experiences to those who work remotely and gamers looking to play untethered from a PC or console," said Jen-Hsun Huang, president and chief executive officer of Nvidia.

Nvidia VGX: One Professional Board Can Serve Multiple Users

Nvidia VGX platform enables workers for the first time to access a GPU-accelerated desktop similar to a traditional local PC from any device (thin client, laptop, tablet or smartphone) regardless of its operating system, and enjoy a responsive experience for the full spectrum of applications previously only available on an office PC. The platform's manageability options and ultra-low latency remote display capabilities extend this convenience to those using 3D design and simulation tools, which had previously been too intensive for a virtualized desktop.

Nvidia VGX uniquely addresses the user experience issues of traditional virtualized desktop infrastructure by adding a fully virtualized GPU board to the data center. This now enables up to one hundred users to share a VGX board with a true PC experience with GPU-accelerated VDI (GPU-VDI). Initially, Nvidia will offer one quad-GPU graphics card with 768 stream processors and up to 4GB of memory per chip. The card will be 10.5” in length and 4.4” in height, hence, it will require new cases, special power supply units and so on. Still, Nvidia promises that machines with Nvidia VGX hardware will fit into common server racks.
One of the key features of GPU-accelerated desktop virtualization cloud platform is Nvidia’s sophisticated VGX Hypervisor that manages the GPU resources to allow multiple users to share common hardware, improving user density on a single server while providing true PC performance and compatibility.

Integrating the VGX platform into the corporate network enables enterprise IT departments to address the complex challenges of BYOD (bring your own device) trend and delivers a remote desktop to these devices, providing users the same access they have on their desktop terminal. At the same time, it helps reduce overall IT spend, improve data security and minimize data center complexity.

Nvidia GeForce Grid: High-Quality Gaming Now Available Everywhere

Nvidia GeForce Grid cloud gaming platform allows gaming-as-a-service providers to stream next-generation games to virtually any device with lower latency, while incurring lower operating costs, particularly related to energy usage.
Gamers benefit from the ability to play the latest, most sophisticated games on any connected device, including TVs, smartphones and tablets running Apple iOS and Google Android, provided that game designers implemented special ways to control their titles from those devices.
The key technologies powering the new platform are Nvidia GeForce Grid GPUs with dedicated ultra-low-latency streaming technology and cloud graphics software. Together, they fundamentally change the economics and experience of cloud gaming, enabling gaming-as-a-service providers to operate scalable data centers at costs that are in line with those of movie-streaming services.

The first GeForce Grid graphics card will feature two GPUs each with its own encoder and 1536 stream processors. The boards enable providers to render highly complex games in the cloud and encode them on the GPU, rather than the CPU, allowing their servers to simultaneously run more game streams. Server power-consumption per game stream is reduced to about one-half that of previous implementations, an important metric for data centers.
Fast streaming technology reduces server latency to as little as 10 milliseconds by capturing and encoding a game frame in a single pass. The GeForce Grid platform uses fast-frame capture, concurrent rendering and single-pass encoding to achieve ultra-fast game streaming.
The latency-reducing technology in GeForce Grid GPUs compensates for the distance in the network, so gamers will feel like they are playing on a gaming supercomputer located in the same room. What remains to be seen, though, is whether such encoding and decoding of a very dynamic video stream will provide the same level of quality as local PCs.
Nvidia and Gaikai demonstrated a virtual game console, consisting of an LG Cinema 3D Smart TV running a Gaikai application connected to a GeForce Grid GPU in a server 10 miles away. Instant, lag-free play was enabled on a highly complex PC game, with only an Ethernet cable and wireless USB game pad connected to the TV.

"A cloud GPU is transformational technology for the online gaming industry. Using a graphics-optimized cloud GPU eliminates some of the final barriers that stand in the way of a truly immersive, truly exciting cloud-based gaming experience and could change the business model for how games are played and delivered," said Brian Blau, an analyst with Gartner.