In mid-2016, Intel announced that it will work with BMW (Germany) and Israel Mobileye to jointly develop autopilot vehicles, and the product will be available in 2020. Missed the era of mobile Internet, Intel will be the field of Mo’s Law then into the era of autopilot? Although doubt, but also worth the wait, at least the influx of competitors, the industry’s future clearer.

Intel, which is believed to have missed out on the mobile Internet era, has recently turned its attention to the field of autopilot, but Intel does not seem to be too early in this year’s hot autopilot. Fairchild Semiconductor

It is reported that semiconductor giant Intel (Intel) ready to invest heavily in the field of autopilot cars, the development of autopilot dedicated car processor.

Moreover, the current Intel has been set up within the automotive solutions division, began to carry out related research and development plans.

The report notes that in the past, the development of personal computers, Intel has almost become a personal computer processor, the word.

However, with the gradual decline of the personal computer industry, Intel also had to transition to other areas of development. In the upcoming era of autopilot, Intel also hope to copy the success of personal computer processor experience.

In fact, the autopilot is actually a computer-controlled car. Therefore, the autopilot vehicle’s processor chip must be able to handle a large amount of information coming from the sensors, cameras, and so forth of the self-drive vehicle to analyze and deal with the current traffic, surrounding environment, and road anomalies. Steering wheel, brake, fuel and other systems to implement the appropriate control. skyworks

Intel’s Automotive Solutions Group, Elliot Garbus, general manager of a recent interview, said the self-driving on the intensive data processing business, its strength is similar to the current high-performance servers used in the processor.

So, in essence, an autopilot is a data center that travels on wheels with wheels on land.

Elliot Garbus also stressed that the future of autopilot dedicated car processor in addition to strong processing performance, Intel is also prepared to allow the processor with diversity and adaptability.

A few days ago, Intel’s global vice president and president of China, Yang Xu in Beijing, said that by 2020, with the needs of all things network amplification, sensors and hardware equipment to increase, only a car will produce the data will reach 4000GB / day , Coupled with other areas more large, based on this, we are bound to usher in the data outbreak era.

As in the past, personal computers, automotive processors can also be applied to a wide range of vehicles. In this regard to observe, in the eyes of Intel, auto-driven car has become a personal computer, smart phone, the other a large number of processor chip applications.

Therefore, Intel hopes to accelerate development, to seize market share.

Over the past period, Intel has been completely in the smart phone processor market fall, the market hand over to the American Qualcomm (Qualcomm) and Taiwan MediaTek.

Today, in the autopilot segment, it is clear that Intel has lagged behind Google, Uber, Tesla, and many of the traditional large-scale depots in terms of car-driving systems.

Therefore, Intel in this area, began to focus on the direction of transfer, the past is good at the processor development.

In mid-2016, Intel announced that it will work with BMW (Germany) and Israel Mobileye to jointly develop autopilot vehicles, and the product will be available in 2020.

In this partnership agreement, Intel plans to focus on the development of good processor chip, and Israel Mobileye company in computer image recognition, machine learning, and data analysis and other fields.

In addition, in May 2016, Intel also acquired a computer vision recognition company Itseez, was considered to be for the field of self-driving technology.

Of course, in the field of auto-drive car chip, Intel will also face a large number of strong competitors challenge.

This includes Nvidia, Qualcomm, and Renesas. In fact, in the face of the global PC market decline, Intel has been looking for its own position. Before, Intel has been in the Internet of things chip, UAV chip market and other extensive investment.

However, the size of these markets is still very limited, can not make up for the computer processor chip market decline. Today, Intel hopes to be able to drive a car in the wave of auto-sharing a market to recover the future development trend.

In the Chinese market, yesterday Intel and Red Flag, Neusoft tripartite cooperation announced the development of “smart cockpit”, Intel on October 25 the latest release of Intel Atom car processor (codenamed: Apollo Lake), this car processor to support virtual And seamless support for automotive infotainment systems, digital instrumentation and car navigation systems and other high-definition screen use and interaction.

Of course, Intel also has its unique advantages, in its business focus will be transferred to the data center chip and Internet access device chip in the process, in the data center processor market, Intel has more than 99% market share. At present, only Intel can provide all the components, including hardware, software, storage, network and security technologies to limit the efficiency and reduce costs.

Intel, through its acquisition, has already begun building its automotive portfolio with Wind River to partner with automakers and suppliers, as well as software companies such as Green Hills, QNX and other key partners to deliver software-oriented Define unmanned solutions.

Missed the era of mobile Internet, Intel will be the field of Mo’s Law then into the era of auto-driven auto it?

Although doubt, but also worth the wait. At least the more influx of competitors, the future of this industry more clearly.

Researchers at the University of Illinois have developed a new thermal control method for gallium nitride (GaN).

GaN transistors have higher power densities than conventional silicon transistors and can operate at higher temperatures (below 500 ° C), but GaN transistors, like all semiconductors, also generate excessive heat, which limits their performance.

Cooling methods based on heat sinks and fans add cost and volume. Now, a team of researchers at the University of Illinois’ micro-nanotechnology labs has created a new approach that claims to be simple and low-cost. smbd-101

Using computer-aided design, Kan Bailam’s team has demonstrated that the thickness of the GaN layer plays a large role in overheating, affecting the thermal expectations and final performance of the device.

“Thinner and colder,” says Byram, “traditional GaN transistors are deposited on thick substrates (such as silicon, silicon carbide), but this is not an ideal thermal conductor.” The challenge, he says, Epitaxial mismatches on the substrate, which results in devices that are tens of microns thick, and in many cases hundreds of microns.

“This has a very poor effect on heat dissipation, taking into account the heat source as the gate narrows at the submicron level,” Beylam said. The use of novel semiconductor release methods, such as smart cutting and peeling, a GaN transistor that can be released from the remaining epitaxial and thicker substrates, can benefit from improved thermal control. mt47h64m16hr

“By refining the device layer, you can reduce the high-power GaN transistor 50 degrees Celsius hot spot temperature.” Byram said.

The research leader Kehun Parker believes there is a limit to device thickness. “If you reduce too much, you get the opposite effect, actually increasing the temperature inside the device,” Parker said. For a typical device, the optimum thickness is about one micron.

The work, they say, is important because it provides thermal-design guidelines for GaN-based transistors. The optimal layer size is closely related to the device interface thermal resistance (TBR), which is a condition for the existence of GaN epitaxial layers and other interfaces.

“We determined the optimal thickness of the GaN transistor’s hot-spot temperature, depending on the value of the TBR, and the size of the layer depends on how the device will be used.” If it is for higher power applications, then ideally you need more Thin, sub-micron thick layer. ”

The next step for the Illinois team is to study the electrical properties of GaN layers, such as engineering diamonds and epitaxial graphene, before actually fabricating GaN transistors on substrates.

The Illinois study was funded by the Air Force Academy of Science (AFOSR) Youth Research Project Grant No. FA9550-16-1-0224; the results were published in the October 10, 2016 Applied Physics Letters B1 page.

Global semiconductor capital spending is expected to fall 0.3 percent to $ 64.6 billion in 2016, slightly higher than the 0.7 percent decline forecast for the previous year, according to Gartner.

Gartner said the market is expected to grow by 7.4% in 2017.

“We see a return to growth in the semiconductor industry in the last quarter of 2016,” said David Christensen, senior research analyst at Gartner Inc. “In 2017, wafer manufacturers will focus on introducing high-capacity 10-nanometer products, with memory makers focusing on 3D NAND flash. ”

A few years ago, smart phones, mobile devices, solid state drives and things will continue to lead the semiconductor market development, especially foundries, for these devices to produce most of the wafer chip. hmt41gv7bmr4a-h9

Although the smart phone shipments slowed, but 4G LTE in the field of high-end smart phones led to the rapid development of advanced chip technology needs, and China’s smart phones using fingerprint sensors, touch-screen drive and AMOLEDs full use of 200mm wafer generation Factory 0.18 micron wafer technology.

From the device point of view, the 2016 first half of the DRAM situation is even worse, mid-market bottom. Now supply and increased demand led to the second half of the market supply shortage. Early 2017, weak demand environment will result in over-technology, but then 2017 and 2018, the industry will be in short supply of the phenomenon. QCA150A60

After nearly three years of oversupply, the third quarter of 2016 NAND shortage of obvious, 3D NAND production challenges remain. 2017 is expected to see the balance between supply and demand. The second half of this year, a large number of new capacity and 3D NAND technology matures, will promote the balance between supply and demand.

In 2016, wafer manufacturer’s 10nm product and memory manufacturer’s 3D NAND to promote wafer-level manufacturing equipment spending growth of 6.4%.

“The global semiconductor industry is maturing, with annual growth dropping from more than 10 percent a year to the current 4 to 5 percent per year, driving the industry to maintain a positive M & A trend.” NXP Semiconductors President and Chief Executive Officer Richard De Kelaiming recently in an exclusive interview.

In 2015, the global semiconductor industry mergers and acquisitions more than 120 billion US dollars. Into 2016, and have burst the news of the semiconductor giant mergers and acquisitions, including generous, including Japan’s Softbank spend about 32 billion US dollars acquisition of British semiconductor and software design firm ARM, the United States to 3.56 billion US dollars of micro-technology cash and stock acquisition competition Rival Atmel, Japan’s Renesas to 3.229 billion US dollars acquisition of the United States and Intel. erj3ekf2743v

In the division of thunder, the global semiconductor industry into maturity, the increasingly fierce market competition, the industry will seek to further enhance the integration of mergers and acquisitions, the future there will be an amazing amount of the transaction M & A offer. According to his judgment, the global semiconductor industry will become increasingly centralized trend, the development trajectory may follow the Western military industry, from the initial hundreds of thousands, through constant survival of the fittest, mergers and acquisitions, eventually evolved into a specific area only Supported by several giants.

“In this trend, NXP can not stay out,” Kelaiming said, NXP hopes through mergers and acquisitions, will be more advantages of combining technology to provide customers with comprehensive solutions. For example, NXP completed the end of the acquisition of Freescale, to bring more market opportunities. Data show that the acquisition of Freescale’s new NXP market valuation of more than 40 billion US dollars, annual revenues of more than 10 billion US dollars, became the world’s largest supplier of automotive electronic components. 422lp9r

Transnational semiconductor giants in the active participation in mergers and acquisitions market at the same time, with particular emphasis on deep plowing prospects are generally optimistic about the Chinese market. According to professional research institutions statistics, in 2015 the size of the Chinese market reached 1102.4 billion yuan, becoming the world’s largest consumer of semiconductor products.

Klein once said that NXP in the Chinese market has been the implementation of sound expansion policy, and Datang, millet, the United States and many other Chinese enterprises around the unmanned, mobile payment, intelligent life and other content cooperation. Not long ago, NXP also Changan Automobile, Neusoft Group set up “China Automotive Information Security common interest group” to jointly promote the hardware-based automotive safety industry standard set up and application. This is the NXP in the global market, the first joint local automotive industry, software companies launched automotive safety organizations. Section Lei Ming told reporters that the establishment of this car information security team is critical to NXP. “In the field of unmanned vehicles, the opportunities in the Chinese market are much greater than in other regions, so we have to find the right partner in China to complete the business.” Klein said NXP will provide the Chinese market with Unmanned solution with core technology.

According to reports, NXP is headquartered in Eindhoven, the Netherlands, Philips in 2006 stripped of its own chip and semiconductor business was founded. As the world’s leading semiconductor manufacturer, NXP provides solutions for interconnected automotive, network security, portable and wearable devices and the Internet of Things. With more than 60 years of expertise and experience, NXP has operations in more than 35 countries, with annual revenues of $ 6.1 billion in 2015.

According to foreign media reports, Samsung announced a wearable device for the chip Exynos7 Dual 7270 – the first use of 14-nanometer wearable device application processor.

Samsung in 2015 began production of 14-nanometer smart phone chip. Applied in the international version of Samsung GalaxyS6, S6 edge and Galaxy Note5 Exynos 7420, Samsung is the first 14-nanometer chip, Samsung will now use advanced 14-nanometer process to produce wearable device chips.

The Exynos 7270 integrates two ARM CortexA53 cores – striking a balance between efficiency and performance. Samsung said the Exynos 7270 consumes about 20 percent less energy than its 28-nanometer equivalent chip due to its 14-nm process. hd74ls00p

Interestingly, the Exynos 7270 is also integrated with an LTE Cat.4 modem, meaning that the wearable device configured with the chip transmits data directly from the mobile carrier’s network without the need for a mobile phone. In addition, it also supports WiFi, Bluetooth and FM technology.

Exynos 7270 Another selling point is that it integrates CPU, wireless module, DRAM memory and NAND memory. Samsung said the solution could reduce chip thickness by about 30 percent, which would allow wearable equipment manufacturers to design thinner wearable devices without sacrificing performance or performance.

Samsung did not disclose the first time that the wearable device with Exynos7270 was available for sale, but vendors interested in it already had access to reference platforms – including Exynos 7270 chips, NFC chips and a variety of sensors – to start product design ahead of schedule. lm2594

Although Samsung did not mention, but the industry speculation, Exynos7270 will be the core of Samsung GearS3 smart watches. GearS3 is expected to be released as early as this month.