CEVA Dolphin Weninar SemiWiki 800x100 260419 (1)

AMAT reports strong Q and even better guidance

AMAT reports strong Q and even better guidance
by Robert Maire on 08-21-2016 at 4:00 pm

Last quarter we said that AMAT got its mojo back and it appears to have even picked up speed going into the end of a strong year.

The display business which had been less than reliable in years past has come up with back to back home runs. Applied is growing both its top and bottom line at well above the sluggish market rates and is clearly the growth leader in the industry right now. Its a very nice combination of obvious share gains, good new products, being in the right place at the right time and nice execution. Financials are at an all time high

Driven by conversion inflection points in display, OLED and 7NM OHHHHLED displays…
The inflection point in the display business is the industry conversion to OLED which is running at a rapid pace and will likely drive orders for a while as we are just at the begining of a conversion cycle. Management suggested that OLED was three times larger than prior display business.

3D NAND continues its strength…

The company bragged about a 7 point share gain in 3D NAND with the market for tools up roughly 40% versus last year. This is a very strong performance as Applied hasn’t done as well as competitors in the past in the NAND market but is doing a great job of making up for it in the conversion to 3D. DRAM remains the weakest segment and is off by 25%.

7NM is going to be “HHHUUUGE”….

Applied confirmed what we have heard and previously reported, that 7NM will be a strong node, much stronger than the 10NM node. GloFo is skipping 10NM altogether and we think that the Samsung and TSMC will catch up, if not even pass Intel at 7NM. Logic remains depressed in terms of spend with foundry up about 5-10%.

Blowout Guidance…
Management guided for sales to be up 15 to 19% in the current quarter with semi biz up 10-23%, service up 3-6%, display up a whopping 35-40%. EPS is expected to come in at a record breaking $0.61 to $0.69. All this outperformance in a WFE spending market that is little more up than flattish overall.

New tools should keep growth growing…

Applied is in the early stages of a number of new products that have all had excellent initial success with great ramps. This success is across many market segments and should prove long lived

Gross margin the only soft spot…
The only soft spot on the call were questions about gross margin progress and goals. While progress continues, investors would like to see more faster. Given where we are in product sales and goals of share gain , gross margin may be a bit slow to come around.

Leader of the pack…
Right now, Applied is the momentum leader in the industry , followed by a strong KLAC with LRCX and ASML doing OK but having their own unique issues. This is a significant role reversal from a couple of short years ago from Applieds nadir when it hit rock bottom after the failure of the TEL deal.

Better off alone???
We could easily argue that AMAT is better off not having done the TEL deal as their stand alone performance has been better than we would have expected when weighted down by the baggage of the TEL deal. Maybe the regulatory agencies did them a favor.

#NEVER EUV
The rest of the industry, including Applied, has benefited from the lack of EUV and need for multi patterning. This is not going to change any time soon as Mark Bohr, CTO of Intel, admitted this week at Intel IDF that Intel will not likely use EUV at 7NM, and only use EUV when its ready for prime time which it very obviously isn’t.

This is obviously no surprise at all to us as we have been the most vocal analyst about the delay of EUV, but it was obviously a surprise to some naive analysts who believe the party line as ASML’s stock was down this week on the news.

The Stocks…

A $30 stock price seems in the cards in the near term for Applied. We see no reason to get off the train as there is no slow down in sight. Longer term we may keep an eye out for an eventual display slow down but not for a while.

Applied’s performance should generally be a rising tide for all ships in the semi equipment ocean, but we would remind investors that Applied’s gains are coming from somewhere, like from Lam in etch and Dep, KLAC in E Beam, and ASML in multipatterning.

But investors are not likely that insightful and will probably bid up most stocks, across the board, on the news.


Alphabet vs. Radio: Traffic Info Smackdown

Alphabet vs. Radio: Traffic Info Smackdown
by Roger C. Lanctot on 08-21-2016 at 12:00 pm

Alphabet is showing signs of taking over just about every infotainment function in cars except one: traffic information. When Alphabet, aka Google, arrived on the automotive scene the company and its minions quickly grasped that there was going to be a big opportunity for search, voice, navigation, maps, traffic and contextual marketing and advertising messages. It was also clear that there would be a growing need for more and more software and applications – providing tasty tidbits for Android and its growing developer community.


In spite of the many challenges for a dominant player in the mobile market to shift its content, services and applications to the cost, power and processing constrained automotive environment, Google dove straight in. Within a very short time Google, now known as Alphabet, took charge of in-vehicle search, despite the persistence of CloudMade, Bing (now out of the automotive market) and deCarta (now owned by Uber).

The next step was to provide for the integration of Android-based smartphones for which Alphabet created the Open Android Alliance and launched the Android Auto platform which brought Google Voice, navigation, content and the GoogleNow contextual messaging platform into the dashboards of most of the leading car makers.

In the process, Alphabet, like Apple, began taking charge of in-vehicle integration efforts by introducing a certification program for the deployment of Alphabet services, content and applications. Suddenly car makers discovered that they had handed the car keys to a search company.

Alphabet (and Apple) have now taken over certification of automotive center stacks to ensure that content is rendered with the proper colors, frame rates, resolution and fonts, among other things. This is a bit of a shocking turn of events, but it all comes back to search.

Auto makers seeking to bring search to the driving public realized very early on that Google had the best search. Other than Bing, no player of Google’s size was either interested in or prepared to deliver an automotive grade search experience.

That search dominance cracked open the door for Android Auto and eventually native Android (Android N) in the center stack and possibly elsewhere in the dashboard. But a survey conducted by Jacobs Media points up the lingering Achilles heel of Alphabet: traffic information.

Jacobs Media surveyed some 39,000 listeners to 245 radio stations in the U.S. and Canada in Jan. and Feb. 2016. According to a RadioWorld report, “most respondents were members of station databases, and some replies were gathered via station websites and/or social networking pages. All were collected online and weighed using metro population data.

“The Web survey doesn’t represent all radio listeners or even each station’s audience. Because it’s an opt-in, no margin of error is calculated.

– See more at: http://www.radioworld.com/article/jacobs-scrutinizes-our-changing-media-lives/279206#sthash.trzKPO0q.dpuf”

Key results of the survey are illustrated above the headline to this blog. For radio listeners interested in traffic information, radio remains the dominant and preferred source. The survey does illustrate, however, the rising influence of Google Maps and Waze as resources for traffic info. (The survey does not distinguish between in-vehicle and non-vehicle use – hence, television is yet another strong source of traffic information.)

Second only to search, traffic information is a must-have for the driving public. Strategy Analytics surveys universally show traffic as the single most desired and most frequently used application. It is interesting to note the emergence in the Jacobs Media survey of Waze as a competing source for traffic information.

In the U.S., Waze is to traffic as Pandora is to streaming music. Waze, owned by Alphabet, is an increasingly popular traffic and navigation application, but not yet as popular as radio for traffic information overall.

Just as respondents to the Jacobs Media survey who are interested in traffic data likely prefer a live radio voice reporting the real-time traffic situation, listeners to broadcast music express their strong preference for live DJs and the opportunity for new music discovery in a curated radio environment.
The Jacobs Media survey and a survey conducted by Edison Research working with NextRadio both show the importance of the curated broadcast experience.

Pandora and Waze represent automated experiences trying to deliver contextual experiences to cars via connected smartphones. It looks like a man over machine proposition as listeners continue to choose humans over algorithms. In spite of the ongoing fragmentation of content consumption in the car, radio continues to dominate in frequency and amount of listening.


The slow deterioration of that dominance is clear from Strategy Analytics’ own survey research. Alphabet may be gaining ground in embedding Android, integrating smartphones and propagating Google Search and Voice and Now, but traffic dominance remains elusive.

It was only five years ago that Google initially partnered with INRIX before parting company with the Seattle-based firm. The attraction of INRIX was its predictive traffic model, something also available from HERE and TomTom and SiriusXM. Google and Waze have a reputation for emphasizing real-time traffic information and alerts.

Alphabet’s failure to take over in-vehicle navigation and its inability to establish predictive traffic credibility, has left the company on the outside looking in. This weakness will only become more pronounced as digital radio in the form of DTS’s HD Radio sees wider distribution along with a shift from RDS-TMC traffic information to T-PEG – offering several times the volume of information delivery free over the airwaves.

Alphabet is also facing the amazing competitive proposition of radio becoming a searchable medium thanks to NextRadio technology and the wider availability of radio-enabled smartphones with digital reception. But whether it is traffic or music, radio listeners continue to demonstrate their interest in and willingness to respond to and reward radio stations with live talent curating a local broadcast experience. It’s mankind vs. algorithms and the talent is winning, for now.

Roger C. Lanctot is Associate Director in the Global Automotive Practice at Strategy Analytics. More details about Strategy Analytics can be found here: https://www.strategyanalytics.com/access-services/automotive#.VuGdXfkrKUk


Five Ways IoT is Changing Trends in Cognitive Business

Five Ways IoT is Changing Trends in Cognitive Business
by Bill McCabe on 08-21-2016 at 7:00 am

How is the Internet of Things, or IoT, changing trends in cognitive business? The impact is evident with the alliance between computers and humans. Computers still quickly assimilate and spew out data based on what is seen, heard and read. The difference is that they are now reasoning, understanding, and learning from those processes, whether using older, stored information or real-time data. Swift, accurate decisions are based on facts rather than emotions or circumstances. Here are five exciting business trends motivated by IoT.

1.Intense communication skills: It’s common to hear people wish they could be more like computers. Today’s cognitive systems aspire to be more like humans and they are doing a spectacular job of achieving their goal. Artificial Intelligence can now grasp concepts and emotions including:

  • Environmental conditions
  • Sentiment
  • Personal relationships
  • Emotional undertones

The result is an analysis of untold amounts of data tempered or enhanced by the lessons and emotions learned from interacting with humans. Rather than a strict response, cognitive systems develop a theory based on probability and the effect of unpredictable information.

2. New operations and processes: Cognitive IoT has the power to achieve unimaginable capabilities that change company operations. Internal and external data have additional value because the system continues to learn how to understand and express possibilities of context, workflows, and environments. Processing speed and calculations are far superior to the human brain, while effective forecasting improves as cognitive IoT learns and understands more about data through interaction with artificial and human intelligence.

3. Smart products and services: Financial services and medical devices are examples of business changes due to cognitive systems. Feedback from consumer reactions guides practices such as greetings, sales events, and fit and comfort of medical equipment.

4. Specialized expertise in the workforce: Cognitive systems can be designed to learn and the unique language of any profession. Expertise in the wording of retail, manufacturing, and medical fields allows the system to efficiently teach specialized words in any field to employees, reducing the time required by humans to study and learn.

5. Advancement and innovation: Cognitive technology results in the IoT searching for business opportunities and consumer patterns while processing data. It will provide well-documented suggestions for businesses of any size to consider acting upon. The number of sources and intelligence feeding into the system delivers innovative suggestions for advancement. Humans, programmable systems, and traditional research are unable to accomplish the swift sorting and determination of data. The results, however, are well documented to allow management to make a rational decision about the future of any suggestion.

For more information check out www.internetofthingsrecruiting.com


4 Reasons for Bluetooth 5 Adoption in IoT

4 Reasons for Bluetooth 5 Adoption in IoT
by Eric Esteve on 08-21-2016 at 4:25 am

IoT devices have to be connected but power consumption is usually a real concern. If you think about wearables, like for example fitness wristbands, the time between charges could make or break the product. Even if Wi-Fi looks attractive to connect an IoT device, the system developers have quickly realize that the power consumption associated with Wi-Fi technology was too high, leading to short active time before recharging. Bluetooth 4.0 (aka Low Energy or BLE) has generated high interest from the developers of emerging IoT applications.

But Bluetooth has been initially defined for short range usage, typically headset and smartphone, with paired device broadcasting approach. According with Bluetooth SIG, the future launch of Bluetooth 5 at the end of 2016, beginning of 2017, will not only suppress these limitations in term of range and broadcasting capability, but also double the speed and by consequence half the power consumption. Let’s review in details these four improvements and their impact on Bluetooth 5 adoption for battery powered connected devices.

Doubling the speed means that Bluetooth Low Energy 5 (BLE5) will support 2 Mbit/s data rate, doubling the version 4 (BLE) data rate (1 Mbit/s). For most applications, the speed of current Bluetooth standard for relaying data to the platform (smartphone or gateway) was sufficient. If we take again the example of the fitness wristbands, the amount of data to be transferred is pretty modest and BTLE data rate support was enough. Nevertheless, even for such wearable devices, higher transfer speed will allow faster software and firmware updates and improves user experience.

One consequence of doubling the speed will have a very strong impact on user experience as well: double speed means half the power consumption! In other words, we can expect the time between charges to last longer, up to twice the time when using the previous Bluetooth release. This impact on power consumption would not be true if using wired devices, like high speed SerDes. In fact, if you double the speed of wired device (SerDes on the same technology), you almost double the power consumption. But Bluetooth is a wireless technology and the radio frequency stays the same at 2.4 GHz with Bluetooth 5 and the radio frequency dictate the power consumption, not the data rate. For the same power envelope, Bluetooth 5 allows exchanging twice the amount of data and the consequence is that the device consumes half the power to exchange the same data. The magic of wireless technology!

When selecting a wireless standard for battery powered IoT application, providing the same amount of data at half the power is already very important. The next improvement may have an even stronger impact, as the new standard will increase the range of communication by a factor of four. For existing wirelessly connected consumer devices, Bluetooth standard is ubiquitous, like USB for wired connection. If you need to be convinced about the standard ubiquity, just take a look at Bluetooth SIG web site and you will realize that it references 30,000 member companies. Extending the range of communication by a factor of four will allow users to gain control of their smart home products from anywhere in their home. By the way, this remove the main advantage of using power hungry Wi-Fi to control smart home devices. IoT application will see wide adoption at the express condition that the devices can be affordable at low cost. As of today, many products have multiple and redundant options for connectivity. Bluetooth 5 could become the unique wireless standard if it offers the same benefit than Wi-Fi in term of range or data rate, but much lower power consumption.

The range improvement going with Bluetooth 5 may also open new doors, especially when we consider the boost in broadcasting offered by the new standard: 800% improvement. Increasing broadcast capacity will propel the next generation of “connectionless” services like beacons and location-relevant information and navigation. If we can receive and interact with these services without installing any additional applications and without the nuisance of setting up a connection beforehand, we clearly open up more possibilities for Bluetooth powered systems. The four improvements offered by Bluetooth 5, doubling the speed and half the power consumption, increasing the range of communication by a factor of four and the boost in broadcasting, added to the ubiquity of the standard in consumer devices, should allow wider adoption of the standard, including emerging IoT. It’s even possible that Bluetooth 5 eventually stays the unique wireless standard implemented in many interconnected application…

Thanks to acquisition of RivieraWaves, CEVA has expanded from Digital only DSP and SATA/SAS IP, to support wide interconnect solution, adding WiFi and Bluetooth IP to the port-folio. We have already written in Semiwiki how smart this move was at the time the deal was made. If we take a look at Q2 2016 results from CEVA, the IP vendors completed ten license agreements and six of these agreements were for CEVA connectivity IP. We don’t know how many of these six licenses were for SAT or SAS IP (ranked in connectivity) but we can guess that a majority was for wireless connectivity.

According with CEVA, target applications for customer deployment are vision processing for a virtual reality product (DSP), 5G base stations (DSP), voice processors (DSP), Bluetooth low energy and Bluetooth 5 connectivity for various IoT devices and storage drives (SATA/SAS). Did you notice than in advance of the official standard release by the SIG, CEVA has already closed several deal with Bluetooth 5 for IoT device? This is good sign for CEVA and good sign for Bluetooth 5 adoption in IoT!

We have found many interesting information in this blog “Will Bluetooth 5 Enhancements Make BLE the Best Choice for IoT? from Franz Dugand, Director of Sales & Marketing, Connectivity Business Unit, CEVA.

You can find the Press Release from Bluetooth SIG announcing Bluetooth 5 here

Eric Esteve from IPNEST


What is Inside the iPhone 7?

What is Inside the iPhone 7?
by Daniel Nenni on 08-20-2016 at 7:00 am

TSMC is the bellwether for not only the foundry business, since they are the dominant player, but also the semiconductor industry as a whole. You could also argue that TSMC is a sneak peek into the world economy since they build capacity based on their customer’s forecasts and the world now revolves around semiconductors.

The other thing you need to consider is what we call the “Apple Effect” meaning that as TSMC’s largest customer (both directly and indirectly from the Apple semiconductor supply chain), TSMC’s forecasts are a direct peek into Apple’s future.

With the iPhone 7 coming next month I thought it would be interesting to look at what is inside the iPhone 6s to see what may be coming next. In looking at the iFixit teardown here are the recognizable chips inside:

The iPhone 6s includes the A9 multi core ARM based SoC which was built on both the TSMC 16nm and the Samsung 14nm FinFET processes. This was the first time Apple dual sourced their SoC and now that I know the real reason why, I believe it will be their last. For the iPhone 7 I believe TSMC will get 100% of the business using their new 16FFC third generation FinFET process. The other chips (most of which are TSMC manufactured) inside the 6s include:

  • Qualcomm MDM9635M LTE Cat. 6 Modem (vs. the MDM9625M found in the iPhone 6)
  • InvenSense MP67B 6-axis Gyroscope and Accelerometer Combo (also found in iPhone 6)
  • Bosch Sensortec 3P7 LA 3-axis Accelerometer (likely BMA280)
  • TriQuint TQF6405 Power Amplifier Module
  • Skyworks SKY77812 Power Amplifier Module
  • Avago AFEM-8030 Power Amplifier Module
  • Qualcomm QFE1100 Envelope Tracking IC
  • Toshiba THGBX5G7D2KLFXG 16 GB 19 nm NAND Flash
  • Universal Scientific Industrial 339S00043 Wi-Fi Module
  • NXP 66V10 NFC Controller (vs. 65V10 found in iPhone 6)
  • Apple/Dialog 338S00120 Power Management IC
  • Apple/Cirrus Logic 338S00105 Audio IC
  • Qualcomm PMD9635 Power Management IC
  • Skyworks SKY77357 Power Amplifier Module (likely an iteration of the SKY77354)
  • Murata 240 Front-End Module
  • RF Micro Devices RF5150 Antenna Switch
  • NXP 1610A3 (likely an iteration of the 1610A1 found in the iPhone 5s and 5c)
  • Apple/Cirrus Logic 338S1285 Audio IC (likely an iteration of the 338S1202 audio codec found in the iPhone 5s)
  • Texas Instruments 65730AOP Power Management IC
  • Qualcomm WTR3925 Radio Frequency Transceiver
  • Possibly a Bosch Sensortec Barometric Pressure Sensor (BMP280)

According to reports in the media, the iPhone 7 will not be too much different than the 6s so it is not a stretch to think that the chip suppliers will not change much. There is a rumor that Intel has won some of the modem business but as I mentioned previously I do not think (for technical reasons) it will be a significant portion if at all. Even so, TSMC manufactures the Intel modems so it does not really matter either way.

If you look closely at the TSMC Q2 investor call you will see significant optimism that can be mapped directly to Apple. First let’s look at CAPEX. TSMC is projecting a 20% CAPEX increase in 2016 primarily for 10nm and 7nm (10nm and 7nm share the same fabs as did 20nm and 16nm). This jibes with my belief that Apple will continue to use TSMC for the 2017 and 2018 iPhones. TSMC recently surprised analysts with excellent financial results and an optimistic outlook for 2016 and 2017. TSMC also made the prediction that they will have 70% of the 10nm business in 2017 which definitely supports my TSMC exclusive Apple business prediction.

Bottom line: TSMC is the key manufacturer of chips for the Apple iPhone and iPad products this year and next. If you believe as I do that the iPhone 7 will break Apple unit shipment records due to a strong upgrade demand from the iPhone5 and 5s installed base then TSMC has good reason to be optimistic for 2016 and 2017.


More on HAPS hybrid prototyping for ARMv8 with Linaro

More on HAPS hybrid prototyping for ARMv8 with Linaro
by Don Dingee on 08-19-2016 at 4:00 pm

A few weeks ago we previewed a Synopsys webinar describing how they are linking the ARM Juno Development Platform with the HAPS-80 and HAPS ProtoCompiler environment. I’ve had a look at the archived event and have some additional thoughts. Continue reading “More on HAPS hybrid prototyping for ARMv8 with Linaro”


The Package Assembly Design Kit (PADK)… the start of something big

The Package Assembly Design Kit (PADK)… the start of something big
by Tom Dillinger on 08-19-2016 at 12:00 pm

Integrated wafer-level fanout (WLFO) packaging technology is emerging as a foundation for multi-die solutions. Mobile product applications require focus on both aggressive chip-to-chip interface performance, as well as the final package volume. Traditional multi-chip packages using PCB laminate substrates do not readily provide the performance nor density targets. To address these requirements, WLFO packages need to support both side-by-side and stacked vertical Package-on-Package (PoP) topologies, to minimize interface electrical length. And, the reconstituted, molded wafer needs to support thinning process steps, to minimize total volume.

The conventional process flow for WLFO packaging involves precise placement of known-good die (KGD) face-down on an adhesive carrier, then encapsulating the array of die in an underfill plus molding compound. The adhesive carrier is removed, and the reconstituted wafer provides the new substrate for:

  • the addition of interconnect redistribution layers (RDL) on top of the die
  • (laser-drilled) vias and metal deposition through the molding for stacked connectivity
  • backside grinding
  • backside interconnect redistribution for PoP (optional)
  • BGA bumps
  • separation of the final, multi-die configuration

The disadvantages of this flow primarily pertain to the dedication of KGD prior to the WLFO packaging steps. The defect density of the conventional WLFO process increases the cost of yield loss, as the KGD are encapsulated. For example, there is a risk of yield loss die shiftdue to warpage issues, related to the fabrication of the reconstituted wafer molding and subsequent RDL patterning.

Amkor has developed an alternative process technology to address the yield/cost risks of the conventional WLFO flow. Rather than RDL patterning after molding, Amkor’s SWIFT (and SLIM) technology utilizes a pre-fabricated, known-good RDL structure that has been built on a separate carrier. The SWIFT assembly flow is illustrated in the figure below.

Figure 1. SWIFT assembly flow, highlighting the fabrication of a separate carrier for the RDL layers

The RDL build-up structure on the carrier is used for placement of KGD, with die microbumps for attach. After attach, underfill, and molding encapsulation, the carrier is removed an BGA bumps are attached.

The carrier flatness tolerance is tighter, compared to the conventional WLFO reconstituted wafer. As a result, Amkor’s SWIFT supports finer resolution RDL width/spacing patterning (which requires finer exposure depth-of-focus); and, the improved flatness allows additional RDL layers to be fabricated, a key benefit for today’s high I/O count SoC’s.

As a leading OSAT, Amkor’s interface with the foundry is flexible — either Amkor could fabricate the RDL structure on the carrier (SWIFT) or the foundry can do the processing (SLIM, with tighter RDL pitch fabrication available).

I recently had the opportunity to chat with Ron Huemoeller, VP Corporate R&D, Ruben Fuentes, VP Design Centers, and Curtis Zwenger, Senior Director Advanced Packaging and Technology Integration at Amkor, about their WLFO technology. John Park, Product Management Director at Cadence joined the discussion as well, to describe the latest for WLP support in Allegro, and to highlight results of the ongoing Cadence-Amkor collaboration.

John provided an insightful comment to level-set the discussion, “Advanced WLP is blurring the line between what is an IC and what is a package. Design enablement requires a mixture of IC and packaging design plus manufacturing rules for the fan-out RDL layers.”

Ron highlighted the complexity of the SWIFT and SLIM package designs, “We are supporting heterogeneous die integration, with each SoC potentially having thousands of I/O’s. The die attach and assembly to the SWIFT carrier needs to support a very fine I/O pitch — for example, copper pillars on a 30um pitch, 10 rows deep around the die.”

Ron continued, “An additional complexity factor is introduced by multi-chip packages which need to provide larger power and ground currents. We’ve added the capability to include an intermediate substrate in the packaging flow. This substrate includes both RDL connection layers, plus the option for thicker copper places for P/G distribution.”

Figure 2. SLIM-on-substrate package cross-section

The most intricate requirement for SWIFT technology is the PoP configuration. Ruben highlighted, “For a stacked PoP topology, the SWIFT carrier will include a pattern of tall copper pillars — Through Molding Vias (TMV). After molding and thinning to expose the pillars, additional redistribution wires can be added on the backside for subsequent PoP attach.”

Figure 3. Top RDL connected to TMV copper pillar, for package-on-package design

Given all these design complexities, and especially with the intricate design and manufacturability rules, Cadence and Amkor have released a “Package Assembly Design Kit” (PADK), to accelerate the productivity of SWIFT/SLIM package layout.

John provided additional background, “The widely-adopted Allegro Package Designer has been enhanced to support Advanced WLP Design (v.17.2). Package designers have access to productivity aids to assist with meeting package manufacturability requirements, such as the metal patterns needed for material outgassing and mechanical stress relief. The key feature is the integration of the Cadence Physical Verification System (PVS) to the APD environment. The sign-off flow uses IC fabrication runsets for DRC and LVS verification. The PADK is the Cadence/Amkor kit to enable this signoff-level verification.”



Figure 4. Cadence WLFO design and verification flowThere are unique data formatting requirements to enable running (batch) PVS from Allegro Package Designer. “Standard laminate package manufacturing uses the Gerber format. For WLP, mask making requires GDS-II streamed out from APD.”, John reminded me. “Package design geometries require accurate translation to IC process definitions. Circles, arcs, and trace ends require adaptation to GDS-II. Trace tapering and angled connections need special handling, as well.”

Executing LVS on the package connectivity requires exporting a Verilog (or CDL) netlist from APD, as well as the die placement data plus GDS-II for the RDL connections.

Error markers from PVS DRC and LVS are annotated back to the APD design platform to aid with debug.

What’s next for advanced wafer-level package design enablement? John indicated,“Given the interdependent nature of both IC and package technology design rules, Cadence is working to enable more interactive rule and constraint checking support. Our RAVEL rule definition tool can be used for in-design DRC for Allegro WLP design.”

Ron added, “We view this initial PADK release for SWIFT/SLIM as a first step in bridging the gap for WLP package design enablement. The physical verification runsets are just the first phase. This collaborative framework with Cadence will enable us to expand the scope of the WLP technology information for designers.”

Ron’s comment really rang true for me. The IC foundry industry fundamentally relies upon the PDK releases to customers, to enable design teams to develop their internal methodologies for implementation and tapeout signoff. The rapid advances in (multi-die) WLP technology require a similar level of technology and verification runset design enablement, to accelerate adoption. The PADK release announcement from Amkor and Cadence is an defining step in this strategic direction.

For more information on Amkor’s SWIFT/SLIM technologies, please follow this link.

For more information on Cadence’s Advanced WLP product support, please follow this link.

All SWIFT and SLIM images courtesy of Amkor Technology, Inc.

-chipguy


Foundation IP for Automotive: so Stringent Quality Requirements!

Foundation IP for Automotive: so Stringent Quality Requirements!
by Eric Esteve on 08-19-2016 at 7:00 am

The Automotive IC market is not the largest segment, but is certainly the segments expected to grow with the highest CAGR, with 10.8% from 2013 to 2018, according with IC Insights (January 2015). If you consider the pretty long concept/design to production cycle time (7 years or more) as well as the numerous segments just emerging in production, like Infotainment, Vehicle Networks, Driver Information and Driver Assistance, you can predict that Automotive IC market growth will stay in the double digits up to 2025… and probably further, when the various innovations integrating high semiconductor value will generate additional IC production.
Continue reading “Foundation IP for Automotive: so Stringent Quality Requirements!”


If an Intel 10nm transistor fell in the ARM forest

If an Intel 10nm transistor fell in the ARM forest
by Don Dingee on 08-18-2016 at 4:00 pm

Intel’s news at IDF this week about partnering with ARM for foundry services on 10nm set off some wild speculation. It’s not a surprise that ARM would enable Intel – they’ve worked together before, ARM is an equal opportunity ecosystem partner, and ARM has publicly announced 10nm cores taped out at TSMC. Continue reading “If an Intel 10nm transistor fell in the ARM forest”


Score 1 for IDMs vs Fabless Intel King of the Photonics Hill in Data Centers

Score 1 for IDMs vs Fabless Intel King of the Photonics Hill in Data Centers
by Mitch Heins on 08-18-2016 at 12:00 pm

For those of you waiting with baited breath to see the Fabless community with the likes of Broadcom or Qualcomm challenge Intel in the data center space, you were just dealt a significant blow. Intel debuted their long awaited silicon photonics modules for what they called “lightning-fast” connectivity in data centers. Intel announced they are in volume production of their 100G PSM4 (Parallel Single Mode fiber 4-lane) and 100G CWDM4 (Coarse Wavelength Division Multiplexing 4-lane) transceiver products targeted for use in data communications applications, in particular switch-to-switch optical interconnects in data centers. Both of these products can communicate 100G signals up to two kilometers distance.

This is undoubtedly just the beginning but it is a significant beginning. While the holy grail is to eventually integrate this kind of functionality onto the processors themselves, this is a large step forward as Intel claims that for the first time, these transceivers integrate both electrical and optical components onto a single die, including a wafer-bonded hybrid III-V GaN laser. This technology, should make a for a lower cost and easier to make solution. The news was delivered as part of the keynoted address at the Chipzilla Developer’s Forum being held in San Francisco this week.

Earlier this year at the Optical Fiber Communications Conference (OFC), Robert Bloom, Director of Strategic Marketing and Business Development for the Silicon Photonics Products Division of Intel gave a briefing on Intel’s vision for the future of data center connectivity. In his presentation Robert showed the expected data center spending levels for 100G and 400G interconnects through the rest of the decade growing from $1.2B in 2016 to $5.1B in 2020. The growth is impressive and is not limited to just crossing the spine of the data center. By 2020 it is expected that these connection speeds will be used virtually everywhere, in the racks, across rows, across the data center and even between data centers. To do this the drive to integrate the photonics closer to the processors will pick up speed. According to Bloom’s presentation 400G is just around the corner with the transceivers moving from pluggable components as shown by Ms. Bryant to being embedded on the boards and eventually integrated with the switches and servers themselves.

It was also disclosed that Microsoft is an early adopter of Intel’s new technology and will be using it in their Azure data centers. Microsoft has also started testing field-programmable gate arrays (FPGAs) from Intel’s Altera business group in its data centers. Intel completed the acquisition of Altera in December of last 2015. This is yet another stake in the heart of the fabless guys entering this space. It makes you wonder how long it will be before Intel integrates the speed of the photonics with the flexibility of an FPGAs for Software Defined Networking capabilities. Now that would be a killer app.

The question still remains as to when the fabless guys are going to jump into the fray. Just last week Juniper Networks acquired Aurrion, a photonics IP provider, however one of the comments on that article claimed that it was a bit of a stretch to call this fabless in the photonics space as Juniper were most likely going to be getting silicon from a foundry source that wasn’t going to open its offerings up to the general public. None the less, the pressure is growing and with the move towards integration of photonics closer and closer to the processors, it will only be a matter of time before the chip volumes will reach a level where the fabless guys can’t afford to ignore photonics any longer.