Bronco Webinar 800x100 1
Semiconductor Expert Forum

Posted on

No Turning Back on Autonomous Driving

No Turning Back on Autonomous Driving
by Roger C. Lanctot on 07-03-2016 at 7:00 am
Categories: Automotive

Politicians will tell you that Fridays are reserved for announcements (defeats, resignations, indictments) intended to be ignored or lost in the end of week news sink. In that context, the Friday before the U.S. Fourth of July three-day weekend may be regarded as second only to the Friday before Christmas as an ideal opportunity to bury an unpleasant bit of news.

This is why I found it puzzling that BMW, Intel and Mobileye chose this particular Friday to announce a major new strategic cooperation around autonomous driving. You can find the Webcast for the 10 a.m. event being held in Munich here: http://www.live.bmwgroup.com/2016pk/index.html

In light of the first fatal crash involving Tesla’s Autopilot, the timing of this news event now seems oddly prescient. With a little luck, Tesla’s and Mobileye’s most virulent critics will be off on vacation somewhere and unavailable to rain on the autonomous driving parade.

To date, Tesla has proven to be the Teflon-coated auto maker. Tesla’s have caught fire, been shredded in crashes, and seen in-the-field hardware retrofits and significant over-the-air software updates to fix flaws big and small or add or enhance features and functions – while the company has danced between the regulatory raindrops avoiding high profile recalls or even a sales cease and desist order. (Of course, many U.S. states do not allow the sales of Tesla vehicles, most notably Michigan.)

The latest apparent autopilot failure will surely be parsed and analyzed by no less than the U.S. National Highway Traffic Safety Administration, which has initiated an investigation, and Tesla itself. Tesla has not only been fairly Teflon-coated, it has also been fairly transparent, but this trial will be the ultimate test. A life was lost for the first time.

The sad reality of the situation, aside from the loss of life, is that this latest development is likely to cause some safety advocates in the industry to hit their own personal emergency brake. Rather than seeking a deeper understanding of what went wrong – sensor failure, software failure, sunglare, driver distraction – they will insist that Tesla was in the wrong and that the time has come to shut down all of this self-driving nonsense.

Let’s be clear about one thing. The cat is out of the bag. The horse is out of the barn. The autopilot is on the road. Advanced safety systems have saved lives and are saving lives but drivers are still obligated to pay attention and remain in control. Even Tesla’s require a hand on the steering wheel at least part of the time – meaning the driver is expected to pay attention and participate in the driving task even in autopilot mode.

Regulators and researchers are fond of blaming drivers for 90% of crashes, and yet we all want humans to remain in the driver’s seat paying attention. So maybe it’s time we stopped blaming drivers and start trying to better understand how to better help drivers be better at what they do so well. The reality is that it is human drivers that will and are teaching the machines how to drive “better” than the humans.

This BBC report includes a link to a video recorded by the driver killed in the crash showing how Tesla’s autopilot prevented a collision with a truck: http://www.bbc.com/news/technology-36680043

It is life-saving performance like that that likely contributed to a sense of over-confidence in the system, perhaps.

Let’s see what BMW, Intel and Mobileye have to say today about their cooperation – and let’s avoid the hysterical reactions that might cause us to turn away from the substantial progress that has been made to-date along the technological path to universal collision avoidance. Tesla now faces its greatest test of transparency. In the process we will all learn a little more about the technology that promises over the long term to steadily reduce the 1.2M deaths suffered every year on roadways around the world. It is a turning point for the industry, but there is no turning back from the pursuit of safer driving and safer cars.

Posted on

Semiconductors out of step with electronics

Semiconductors out of step with electronics
by Bill Jewell on 07-02-2016 at 7:00 am
Categories: Foundries, Semiconductor Intelligence, Semiconductor Services

The global semiconductor market has been on a decline (three-month-average change versus a year ago) since July 2015 according to World Semiconductor Trade Statistics (WSTS). Although numerous factors affect the semiconductor market (capacity and utilization, prices, inventory levels) in the near term, the long term growth is driven by the growth rate in electronics. The chart below shows three-month-average change versus a year ago for electronics production by key regions and for the global semiconductor market.

China is the largest producer of electronics. Its growth rate has decelerated from the 12% – 14% range in 2014 to the 10% – 13% range in 2015 and to the 8% – 9% range in the first four months of 2016. The slowdown in China electronics production has been fairly moderate. U.S. electronics production was in decline for most of 2014 through April 2015. In the last year U.S. growth has been positive, reaching a peak of 7% in November 2015. Japan’s electronics production has been extremely weak, with declines almost every month over the last two years. Although Japan electronics is a drag on the global semiconductor market, Japan is only 9% of the market. Europe monthly electronics production data is not available, but overall industrial production has been positive in the range of 1% to 3% over the last two years.

The chart below shows three-month-average change versus a year ago for electronics production in the key countries of Asia. Asia Pacific accounts for about 60% of the global semiconductor market. China is over half of Asia Pacific. Taiwan has seen generally declining production in the last few years as much of its manufacturing shifts to China. South Korea production growth turned positive in September 2015 after a year of negative change. Singapore turned positive in March 2016 after a year of declines.

Malaysian growth has remained positive in the last few years. Two key emerging countries, Vietnam and India, are currently growing at over 20%. Over the last year, the Asia Pacific semiconductor market has slowed from 7% growth in April 2015 to declines in December 2015 and January 2016. After returning to positive growth in February and March of 2016, the Asia Pacific semiconductor market dropped 4.2% in April 2016.

The global semiconductor market decline in the range of 5% to 6% over the last five months is out of line with electronics production. The Asia Pacific semiconductor market has been stronger than the global market, but is still out of step with electronics production. The short term factors of semiconductor pricing (especially in volatile memory business) and inventory level adjustments are likely driving the semiconductor market decline. Assuming electronics production levels remain positive through 2016, the semiconductor market should turn positive in the second half of 2016. We at Semiconductor Intelligence are holding to our May forecast of 1% growth for the semiconductor market in 2016.

Posted on

21 months lining up OPNFV-on-ARM for telecom

21 months lining up OPNFV-on-ARM for telecom
by Don Dingee on 07-01-2016 at 4:00 pm
Categories: Arm, IP

Telecom infrastructure is one area where X86 architecture hasn’t dominated historically. Infrastructure gear is spread across MIPS, Power, and SPARC architectures, with some X86, and a relative newcomer: ARM, already claiming 15% share. That’s a stunning figure considering only a bit less than 5 years ago Continue reading “21 months lining up OPNFV-on-ARM for telecom”

Posted on

A Brief History of Platform Design Automation

A Brief History of Platform Design Automation
by Daniel Payne on 07-01-2016 at 12:00 pm
Categories: EDA, Platform Design Automation

Two weeks ago I spoke on the phone with Albert Li, Founder and CEO of Platform DA about his EDA company. Prior to founding Platform DA in Beijing, Li worked at Accelicon which was acquired by Agilent in December 2011. Mr. Li graduated from Tsinghua University and Vanderbilt University, both in Electrical Engineering, and has written over 20 technical papers. His team of engineers are experts at transistor device modeling, cell libraries and creating PDKs (Process Design Kit) for use by foundries, IDMs and circuit designers. They also have branch offices in Shanghai and Taiwan Hsinchu. Their web site at www.platform-da.com features both English and Chinese content.

Related – EDA mergers continue: Accelicon acquired by Agilent

Products

There are three main EDA products and the first one is for device modeling and QA called MeQLaband it can be used for applications like:

  • Device modeling for FinFET and planar devices
  • Statistical modeling and mismatch
  • High voltage device modeling, sub-circuit modeling
  • Built-in modeling library and model card QA
  • SRAM modeling
  • Noise modeling and circuit analysis
  • Design or process optimization


MeQLab, design or process optimization

PQLab is a tool for automating the QA of PDK libraries, saving engineering time and can be applied to:

  • Foundry PDK developers needing to QA a PDK
  • IC designers verify that a foundry PDK meets their requirements
  • IC designers compare two or more PDKs


PQLab

For 1/f noise measurements and characterization they have the NC300 system to apply at the wafer level, device, circuit or even with sensors.


NC300


NC300 measurement results

Services

The three categories of services include:

  • Modeling
  • PDKs
  • Semiconductor IP (Standard Cells, SRAM – compilers)

Engineers at Platform DA can design a test chip for you, perform the device modeling, create a new PDK and even provide you with a standard cell library. Device modeling services cover a wide variety of models:

  • BSIM4, BSIM-CMG/IMG, PSPS, BSIM6
  • BSIMSOI
  • Scalable inductor, transformers, baluns
  • HiSIM, HiSIM-HV
  • GP, Mextram, HiCUM, VBIC
  • Diode – Level 1, 2, 3
  • RF Models
  • Reliability models
  • Statistical models
  • Noise model

PDKs can be created to serve your exact needs:

  • FinFET
  • Logic, Mixed Signal, RF CMOS
  • High speed, low power SOI
  • Scalable High Voltage (LDMOS, BCD)
  • PDK enhancements

If your design needs to work in a hardened environment then considering using the RHBD (Rad Hard By Design) service offered by Platform DA where special purpose EDA tools are used:

  • RadEx – environment-aware model extraction
  • SERSim – Chip level SER simulation for combinational logic circuits

The PQLab tool will be used to automate the QA of your new PDK with:

  • Validation of CDF parameters and callbacks
  • Automatic generation of test patterns for DRC and LVS
  • Pre-layout and post-layout simulation QA and comparison
  • Scalability validation of the correlation of device behavior with layout parameters
  • Circuit validation with a combination of Pcells
  • Layout parasitic validation


Summary
Platform DA has successfully served the Taiwanese and Chinese markets for the past four years, and is now growing into several new geographies: South Korea, Japan, Europe and the USA. Their founders have deep industry experience in device modeling, cell libraries, PDKs and services to tie it all together for foundries, IDMs and circuit designers. Expect to start hearing more news coming from Platform DA and their happy semiconductor design customers.

Posted on

Brexit and Semiconductors

Brexit and Semiconductors
by Daniel Nenni on 07-01-2016 at 7:00 am
Categories: Automotive, IoT, Mobile

Interesting news last week with 51.9% of British voters saying yes to Brexit (exiting the European Union). What does it have to do with semiconductors? Plenty! After reading the media’s take on the subject and talking to friends (experts) in China, Taiwan, and Hong Kong, I must say that there is not a consensus to be found and there probably won’t be until the history books document it years from now.

Most compare Brexit to the Donald Trumpism that is sweeping America (angry but largely misinformed people acting out). But to me the risk is FUD (fear, uncertainty, and doubt) that will be used by devious people for their own personal and political advantage, absolutely. China is most certainly one of the countries to watch during Brexit, Russia as well, due to their focused political strength.

According to my friends, the EU is like a three legged stool with France, Germany, and the UK being the legs. Amongst those legs, France and Germany control the EU semiconductor business (STM and Infineon for example). The EU is also China’s top trading partner. If you take away Britain, which is China’s dominant trade partner in the EU, France and Germany are now a two legged stool under the weight of a country (China) with the fastest growing economy and the most FUD capable government the world has ever seen. China has also made semiconductors their National Charter which is an important piece of this Brexit puzzle (for me anyway).

Take the automotive semiconductor industry for example, which today is a $28B market. Three of the four top vendors are from the EU (NXP, Infineon, and STMicro). The market is expected to grow over the next five years at a CAGR of 5.8%. Most of this growth is in Asia and specifically China. The mobile (smartphone and wearables) and Internet of Things (IoT) semiconductor markets are also examples of industry revenue growth domination by China. The smartphone market is still growing at a 6% CAGR and the IoT chip market is estimated to grow at a 11.5% CAGR. Again, Asia is the dominant market growth for both smartphones and IoT and China is aggressively pursuing these segments as both a consumer and supplier.

In fact, according to a recent survey by SEMI.org, more than half of the new semiconductor fab starts in 2016 and 2017 are in China:


From what I understand, all EU trade agreements will have to be renegotiated and with China under political pressure to continue its record economic growth you can expect heavy handed trade negotiations with Britain, the EU, and the rest of the world. So really, Brexit could not have come at a better time to refuel China’s economic growth engine. Taiwan will also benefit as China’s number one semiconductor manufacturing partner.

Meanwhile, the United States, Japan, Korea, and the other economic super powers are at an all-time low in regards to trade negotiations strength so the Brexit winners here will be China and Russia, my opinion. That’s if Brexit actually happens of course, which is a whole different discussion, one that I have not had yet.

Posted on

Synopsys Launch EV6x Vision Processor… Challenging the Competition!

Synopsys Launch EV6x Vision Processor… Challenging the Competition!
by Eric Esteve on 06-30-2016 at 4:00 pm
Categories: IP, Synopsys

When Synopsys bought Virage Logic in 2010, ARC processor IP was in the basket, but at that time ARC processor core was not the most powerful on the market, and by far. The launch of EV6x vision processor sounds like Synopsys has moved ARC processor core by several orders of magnitude in term of processing power. EV6x deliver up to 100X higher performance on common vision processing tasks than the previous generation EV5x! In fact, even if EV6x is part of ARC CPU IP family, this vision processor is a completely new product, defined to address high throughput applications such as ADAS, video surveillance and virtual/augmented reality.

EV6x has heterogeneous multicore architecture including one to four high-performance vision CPU cores. Each vision CPU includes a 32-bit scalar unit and a 512-bit wide vector DSP and can be configured for 8-, 16- or 32-bit operations. In the maximum configuration of four vision CPU cores, the processor delivers up to 620 GOPS/s with full scatter-gather and predication capabilities. From the competitive comparison table below, we can extract the main EV6x differentiators: at 500 MHz, vector DSP delivers 384 GOPS, or 50% more than the closest competitor; CNN performance is even more impressive, with 800 MACs per cycle, when the other two competitors are offering 128 MACs/cycle!

Why does Convolutional Neural Network (CNN) becoming key part of a vision processor? Because CNN is supporting deep learning and this approach outperforms other vision algorithms. Attempting to replicate how the brain sees, CNN recognizes objects directly from pixel images with minimal pre-processing. For vision, CNN can be used to process multiple tasks, like image classification, search for similar images, or object detection, classification and localization. These tasks are supporting automotive ADAS systems, for example, but not only. EV6x vision processor will support surveillance application as well as drones, virtual or augmented reality. CNN is very efficient in face recognition, visual attention and facial expression or gesture recognition.

The CNN engine integrated in the EV6x vision processor not only delivers 6x higher performance than competitive solutions, it also represents the state-of-the-art power efficiency, offering 1000 GOPS per Watt. The CNN engine operates in parallel with Vision CPU, increasing efficiency and throughput. Such performance allows running real-time, high quality image classification, object recognition or semantic segmentation. Last point, this CNN engine is fully programmable to support full range of fixed point CNN graphs, and supports resolution up to 4K.

EV6x vision processor being part of Synopsys DesignWare has been designed for easy integration with any host processor (ARM, Intel, ARC…) in a SoC as the embedded vision processor supports AXI interconnect. Synopsys is used to deliver a complete design environment, offering HAPS FPGA platform for hardware prototyping, reference design and virtual platform, allowing early software development, using EV6x simulation models.

The automotive as well as the surveillance market are very dynamic these days and fast Time-to-Market (TTM) is becoming critical. Synopsys offers reference designs for speed sign or face detection and has built partnership with vendors delivering video surveillance or ADAS application software. The company has made the porting of OpenCV 2.4.9 to RISC scalar core and of OpenVX 1.0.1 standard kernel library, both standard codes being dedicated to embedded vision application.

Availability

The DesignWare EV61, EV62 and EV64 processors are scheduled to be available in October 2016. The MetaWare Development Toolkit and EV SDK Option (which includes the OpenCV library, OpenVX runtime framework and OpenCL C compiler) will be available in June 2016.
Learn more about Synopsys’ Embedded Vision Processors: DesignWare EV6x Processor Family

From Eric Esteve from IPNEST

Posted on

Software Security is Necessary but NOT Sufficient

Software Security is Necessary but NOT Sufficient
by Dr. Jason Oberg on 06-30-2016 at 12:00 pm
Categories: IoT, Security, Tortuga Logic

As the silicon designs inside the connected devices of the Internet of Things transition from specifications to tapeouts, electronics companies have come to the stark realization that software security is simply not adequate. Securing silicon is now a required, not optional, part of RTL design processes.
Continue reading “Software Security is Necessary but NOT Sufficient”

Posted on

From Zero to IoT Prototype in One Month

From Zero to IoT Prototype in One Month
by Bernard Murphy on 06-30-2016 at 7:00 am
Categories: Arm, Cadence, IoT, IP

The best things in life may not always be free, but they don’t have to be incredibly difficult to get to. A challenge for IoT designers has been that their bubbling excitement over the potential of their new gizmo is quickly tempered by the complexities of actually building the hardware. Not exactly what they have come to expect in a Maker world with all kinds of pre-proven open-source software.

But now ARM and Cadence have shown it doesn’t have to be that way. They recently collaborated to deliver an FPGA-based reference development board. The system includes interface IP from Cadence and the IoT processor subsystem from ARM. From Cadence there is MIPI SoundWIre, Quad-SPI, I2C and SPI. The ARM IP includes a Cortex-M processor, power management, flash cache, flash controller and the Cordio radio. And of course ARM mbed provides the software foundation to manage all of this.

That leaves the solution developer to add sensors, actuators and wireless functionality through provided expansion ports, hardware customization through the FPGA and whatever value-add software applications they feel they need. The development board includes a touch screen with image generation, so the prototype can be completely self-contained (maybe with addition of a speaker and microphone).

One engineer (possibly from ARM, reference material is not clear on this) was able to put together a working IoT sensor prototype in just one month starting from this platform. That’s the kind of turn-around time developers need in a competitive IoT market.

It’s easy to forget in a system like this that a lot of the value in the development system is the pre-proven integration between ARM and Cadence components – the majority of the system, to which you are adding your own special sauce. This isn’t something easy to forget when you put systems like this together yourself. There’s a lot of back and forth discussion with IP providers and anxiety over whether you’re really using IP as expected and whether you’re effectively testing for all possible corner cases.

Much of this is simply lack of enough insight into and experience with those IPs. When IP vendors collaborate to prove out subsystems, as ARM and Cadence have done in this case, most of those problems disappear – they’ve already been considered and resolved by the people most expert in the requirements of their IPs.

That’s really why you can get to a proven prototype in a month. It’s not just that you’re using pre-designed and validated components. It’s that you’re using a pre-designed and validated subsystem where potential issues between different providers of components have already been resolved. That’s worth a lot. I’m betting most of us would like to see more solutions like this. You can learn more about the Cadence/ARM prototype HERE.

More articles by Bernard…

Posted on

HBM controller IP holds the key to bandwidth

HBM controller IP holds the key to bandwidth
by Don Dingee on 06-29-2016 at 4:00 pm
Categories: IP, Open-Silicon, Synopsys

We were waiting to see what a different roster including SK Hynix and Synopsys would have to say on HBM in the latest Open Silicon webinar. This event focused on HBM bandwidth issues; a packaging session on 2.5D interposers was promised for a future webinar. Continue reading “HBM controller IP holds the key to bandwidth”

Posted on

Circuit Simulation Panel Discussion at #53DAC

Circuit Simulation Panel Discussion at #53DAC
by Daniel Payne on 06-29-2016 at 12:00 pm
Categories: EDA, Events, FinFET, Synopsys, TSMC

Four panelists from big-name semiconductor design companies spoke about their circuit simulation experiences at #53DAC in Austin this year, so I attended to learn more about SPICE and Fast SPICE circuit simulation. I heard from the following four companies:
Continue reading “Circuit Simulation Panel Discussion at #53DAC”