Synopsys IP Designs Edge AI 800x100

Quality in Design Formats has become a must!

Quality in Design Formats has become a must!
by Daniel Nenni on 06-01-2013 at 6:00 pm

Fractal Technologies is a privately held EDA company with offices in San Carlos, California and Eindhoven, the Netherlands. The company was founded by a small group of highly recognized EDA professionals. The scope of Fractal Technologies is to check consistency and validate all different data formats used in your design and subsequently improve the Quality of your Standard Cell Libraries, IO and IP. Fractal Technologies offers Crossfire software as well as services and customization.

Rene Donkers and Johan Peeters are the gentlemen behind Fractal, guys who I worked with at Sagantec years ago, so it was great to catch up with them during this interview:

Q: What are the specific design challenges your customers are facing?

Our customers are facing the constant increase of design complexities combined with multiple design teams working on the same design, all using best of class EDA tools from different suppliers. In short, a huge challenge for Quality Assurance!

How can we be sure that during design all design formats, from schematic to Verilog, Liberty and physical layout give you a consistent representation of your design?

Ask any Design Engineer and he will tell you that checking Design Formats is becoming more and more complex. Data sizes grow exponential and the need to go to smaller geometries will make this problem even more complex.

Q: What does your company do?
Our company checks consistency and validates all different data formats used in your design and subsequently improve the Quality of your Standard Cell Libraries, IO and Hard IP.

Q: Why did you start/join your company?
I have started the company beginning 2010 as a spin off from Fenix Design Automation together with 2 ex Sagantec colleagues. We truly believe that validation of design formats should be done by an independent tool provider. Internal developed solutions also work but this is definitely not core business for Design groups or CAD teams.

Started my carrier at Sagantec beginning of the 90ties and co-founded Fenix Design Automation in 2006. Becoming CEO of an EDA company has never been on my “to do” list but turns out to be 1 of the interesting challenges that can happen when you start a company.

Q: How does your company help with your customers’ design challenges?
Our tool, Crossfire, provides a vendor independent, automated solution for validation of Consistency and Quality of the Design Formats. Crossfire will read and cross-check all the various formats and views like Open Access layout & schematic views, Milky-Way database, Verilog, Tetramax, VHDL, Liberty .lib, Lef, Def, GDSII, Oasis, Spice, Spectre, Spef(beta), Fastscan/ATPG, STIL/CTL (Core Test Language), HTML, documentation.

Whatever information is provided in a certain format, like delay paths in a .lib file, must be consistent with all other formats. Crossfire is the most complete tool in the industry for checking the quality of designs. The API in Python, Perl and TCL allows adding more checks in a fast and simple way.

Q: What are the tool flows your customers are using?
Crossfire users are split into 3 main groups:

  • Used as Signoff for:

    • Standard Cell / IO Libraries
    • Hard IPs

  • Used in Design Flow in:

    • Library Groups
    • Characterization Groups
    • IP Groups

  • Used for Incoming Inspections for:

    • Standard Cell Libraries
    • Hard IPs

Q: What will be the focus at the Design Automation Conference this year?
At DAC focus will be on the latest checks and features in Crossfire for Hard IP validation, for example support for Spectre, .lib.gz., waiving mechanism and the HTML reporting capabilities.

Q: Where can SemiWiki readers get more information?
All information on our company is on our website:
www.fract-tech.com
Interesting to read is the Crossfire White Paper:
White paper Crossfire
and, if you are at DAC and interested in our products please visit our booth:
Fractal at DAC 3-5 June 2013

lang: en_US


The Ugly Stepchild of Physical Verification – Thermal!

The Ugly Stepchild of Physical Verification – Thermal!
by Daniel Nenni on 06-01-2013 at 3:00 pm

Thermal analysis has traditionally been given short shrift when compared to other more prominent issues facing chip designers. Invarian, to my eye at least, feels that the winds of change are in the air. Not that power or EM/IR issues will fade, that indeed is not the case and in fact quite the contrary, they are contributors to the changing dynamics.

I sat down with Invarian and discussed two test cases, the high and low, so to speak. The first case took us deep inside the bowels of a very large and complex IC. The second case involved the workings at the macro level in dealing with stacked-die configurations. These two cases were chosen by Invarian to demonstrate that thermal analysis is slightly off-center (not unlike our home planet) when compared to the usual suspects in physical verification.

Invarian’s philosophy seems to be that an integrated thermal tool must work from soup to nuts (or the egg to the apples, as the Romans used to say). In other words, to span from the chip world to stacked-die and complex package configurations, all within a single framework. That is Invarian’s value proposition. It’s an interesting approach and certainly a different path than a traditional path for thermal analysis. As for being better, I’ll let the electrons do the thinking for me!

Case A) Thermal Transient Headaches for Network Processor Designs

Invarian’s physical analysis tools live at the tail-end of design simulation. They must bridge the gap between generic corner-based electrical simulation and physical reality. Network processors also live at the edge. They are very large high-performance devices designed at leading-edge processes. There is a general consensus that thermal and transient issues will play a larger role as process technology scales downward. What surprised me, not to mention the designers, was that catastrophic thermal transient effects were discovered at the 28nm node. The question is, was this a fluke or a harbinger on things to come?

Working side-by-side with the physical verification team, Invarian was able to demonstrate that issues of this sort get worse. Their AEs have done extensive analysis at 28nm, and have begun 20nm, to specifically analyze the thermal component and it’s effects. The conclusion was that only by taking a holistic approach where thermal analysis was part of an overall plan were errors caught and corrected. Invarian tools provided a framework for concurrent analysis, taking all the effects of power, timing, EM/IR and thermal, which they felt was necessary to get a handle on what they believe will become a major factor for many designs of the future.

Case B) True 3-D Efficient Thermal Simulation for the Micro and the Macro World

Invarian claims that their 3D thermal tool scales from sub-transistor levels to complex stacked-die configurations. Working with CAD engineering and packaging groups to solve 3D electro-thermal and mechanical stress challenges, they believe, requires an ability to scale up for gathering data, scale down for pinpoint analysis, an ability to generate ‘real-world’ activity, and to do all this quickly and efficiently. It certainly is a big challenge.

Modeling the total environment is a necessary condition, according to Invarian, for capturing enough data for accurate analysis of electro-thermal transients. And there are no thermal tools, at least that I know of, available on the market that scale from the packaging world down to the TCAD level. An important feature of this tool, in my opinion, is its ability to build an optimal thermal grid for the solver.

A natural progression to 3D transistors, low-k dielectric materials and stacked-die configurations is pushing the market for better 3D thermal analysis tools. What began, as an R&D effort with Invarian’s high-performance and mobile customers, seems to be steadily moving toward design engineering. At its heart, Invarian makes a basic assumption, and that is that physically correct models are necessary to get accurate temperature results. In addition, they believe that the tool must easily transverse between the macro world and the micro world. Activity data can also play a critical role in ‘hot spot’ thermal management. What I have learned is that efficient 3D thermal tools must not only generate accurate results in a timely manner, they must also navigate between the various design and research centers within an organization.

InVar Pioneer Thermal™
provides the industry’s largest capacity and most accurate thermal sign-off analysis available today. Invarian solves the problem of miscorrelation with a unique approach to thermal analysis. Different analysis engines work in concert and take into account the interdependence of power, timing, voltage, and temperature into account. Contrary to other tools, all types of analysis are performed in a continuous temperature/voltage space across the chip. InVar does not use predefined corners for analysis.

Invarian is an Exhibitor at DAC, June 2-6, 2013 – Austin, TX – Booth # 1332. On hand will be experts to discuss all the various aspects of physical verification and give a sneak-preview of upcoming products in the areas of ESD and In-rush.

lang: en_US


ARM @ #50DAC

ARM @ #50DAC
by Daniel Nenni on 05-31-2013 at 10:00 pm

The 2013 Design Automation Conference celebrates its 50th anniversary Sunday, June 2 through Thursday, June 6 at the Austin Convention Center. DAC is the world’s leading technical conference and trade show on electronic design and design automation. DAC is where the IC Design and EDA ecosystem learns, networks, and does business. At DAC, visit ARM to learn about our industry-leading technology and see how our technology fits together to accelerate time to market with less risk. In addition, it’s a chance to show how ARM and our Connected Community® Partners are working together to enable the world’s leading consumer electronics companies to bring innovative, energy-efficient and high-performance devices to market.

ARM Booth, #931
Visit ARM as we showcase ARM technology-based solutions enabling smarter systems through integrated and optimized IP. Technology and program highlights include:Cortex®-A15 processor • big.LITTLE™ processing • POP™ IP • Physical IP • DesignStart™ • GPU Compute with Mali™ GPUs • Fast Models with DS-5™ toolchain • ARM University Program • ARM Accredited Engineer Program • ARM-Powered® Device Playground


Presentations and special activities in the ARM Booth:

  • We’ve got more than 25 presentations in the ARM Theater, learn about our newest product offerings and find out more about our partners’ solutions.
  • Want to win a GoPro Hero3? Visit a demo in the ARM booth and be entered into a drawing for your chance to win. Drawings held Monday, Tuesday and Wednesday at 5:50 pm*.
  • Let us buy you a cup of coffee! At DAC, register for ARM DesignStart, an online portal featuring ARM IP for download, and receive a $10 Starbucks gift card.

* Must have your badge scanned and be present to win

ARM Connected Community (CC) Partner Pavilion, Booth #921

The ARM Connected Communityis a global network of more than 1,000 companies aligned to provide a complete solution, from design to manufacture and end use, for products based on the ARM architecture. Join these partners as they highlight their latest innovations:Adapt IP • Apache • Arteris • ASTC/VLAB Works • Cadence Design Systems • CadSoft • Carbon Design Systems • Lauterbach • Memoir Systems • Mentor Graphics • Mentor Embedded • Samplify • Sonics • Space Codesign Systems • Synopsys • Tanner EDA • Zocalo Tech


Presentations
and special activities in the ARM CC Pavilion

  • Don’t miss your chance to win a GoPro Hero3! Visit a demo at one of the participating partner pods in the ARM Connected Community Pavilion and receive a ticket to enter into a drawing for your chance to win. Drawings held Monday, Tuesday and Wednesday at 5:45 pm.
  • View ARM and partner presentations in the CC Pavilion Theater and be entered into drawings held daily during DAC for your chance to win exciting ARM-Powered devices. Daily drawings to be held at 5:45 pm:

    • Nike Fuel Band (Monday)
    • Google Nexus7 (Tuesday)
    • Kindle Fire (Wednesday)
  • Stop by for refreshments:

    • Afternoon snack: Monday, June 3 at 3:15 pm
    • Cocktail reception: Tuesday, June 4 at 5:15 pm
    • Breakfast: Wednesday, June 5 at 9:30 am

ARM Conference Speakers
Hear from ARM executives and technology experts as they discuss industry and technology trends to help you stay ahead of the competition.

Don’t miss these other exciting activities:
DAC 50th anniversary celebrationDenali party by CadenceSi2 25th Anniversary LunchIP Talks!Other ARM Presentations

Check out our pre-DAC blog:
Kickin’ It Up in Austin at DAC’s 50th with ARM and its Partners

lang: en_US


RFIC Design Challenges at #50DAC

RFIC Design Challenges at #50DAC
by Daniel Nenni on 05-31-2013 at 8:00 pm

RFIC developers used to favor mature silicon processes, typically staying back a couple of nodes behind the leading edge. This bought foundries time for ‘RF-enabling’ their PDKs, and also maximized return on investment for developing RF models and infrastructure IP. Not the case any more, it seems. To address the insatiable need for higher connectivity speeds and wider bandwidths, designers of SoCs with high performance gigabit/gigahertz transceivers will now opt for the latest process node, to benefit from higher transistor speeds and improved power/performance tradeoff. As a consequence, foundries and the IP/EDA ecosystem should now rush to RF-enable the leading process node, so that their customers can meet their time to market constraints.

Backing this trend at 20 and 16nm, Helic recently released a library of resizable, parametric inductor cells that are fully lithography-compliant, meeting the restrictive design rules of these advanced nodes. The increased number of Design Rule Checks (DRCs) at the 20/16nm nodes, threatened to prohibitively increase the design effort needed for inductor design. Helic’s solution completely eliminates the need for tedious and time-consuming manual layout, by fully automating the process of layout creation according to designer specifications. The solution includes the automatic generation of dummy fill in the area taken up by a spiral, for meeting minimum density requirements in the fairly large areas taking up by inductors.

Helic’s parametric inductor cells can be used in a variety of 20/16nm IC and SoC designs, including wireless RF transceivers, multi-gigabit transceivers, frequency synthesizers and clock/data recovery circuits. The company aims to drastically shorten the design cycles of such products, which typically employ a good number of integrated inductors for increased performance (e.g. low clock jitter, wide amplifier bandwidth, etc.).

The library of 20/16nm inductors comprises a variety of spiral geometries, such as square and octagonal, including differential and transformer configurations. All these cells can be easily resized to meet a wide range of inductance, quality factor, operating bandwidth and current handling specifications. Helic also offers inductor compiler tools to further automate layout synthesis and optimization.

Helic will present its solutions for 20/16nm RF design, at the Design Automation Conference in Austin next week (booth no. 1843).

About Helic Helic, Inc.develops disruptive EDA technologies for RFIC and high-speed SoC design. We provide our customers with a comprehensive offering combining design tools, silicon IP and applications support, greatly reducing the development cycles of chips for wireless communications, broadband networking, PCs, tablets and other segments. We provide technology for rapid electromagnetics modeling, RF component synthesis, and signal integrity of silicon ICs and Systems-in-Package. Our solutions have been adopted by several major semiconductor companies since 2000. Helic is headquartered at 2880 Zanker road, Suite 203, San Jose, CA 95134.


SemiWiki Top 10 Must See @ #50DAC List!

SemiWiki Top 10 Must See @ #50DAC List!
by Daniel Nenni on 05-31-2013 at 7:45 pm


This list was compiled by the SemiWiki bloggers highlighting emerging technologies that we have written about and that will be demonstrated at the Design Automation Conference next week. We highly recommend you investigate them further during your time in Austin and please let us know what you think.

Today SemiWiki has more than 35 subscribing companies that we work closely with on white papers, webinars, seminars, conferences, strategic marketing, branding, and a variety of other consulting activities. It certainly has been a mind expanding experience for us all.

According to Google Analytics 610,946 people (IP addresses) have visited SemiWiki since going live in January of 2011. SemiWiki is built on a relational database so we can do data mining in regards to key search terms, trending topics, and demographics. This is all driven by the original content posted on SemiWiki. According to our internal analytics there are 8,582 posts by SemiWiki bloggers and registered members. 934 of those posts are from me. Wow! I need to get a life!

Based on all of the above here is the SemiWiki top ten list:

[LIST=1]

  • iDRMfrom Sage DA. iDRM is a technology that I have worked on for the past 3 years including joint development activities with the foundries and top fabless semiconductor companies. Sign up for a DAC demo here.
  • ACEand AFS Megafrom Berkeley Design Automation. BDA’s customer list includes the top semiconductor companies and foundries all of which helped drive this new product development (I work with BDA as the foundry liaison). Sign up for a DAC demo here.
  • High Sigma Monte Carlo from Solido. You can sign-up for the DAC tutorial (featuring me): Winning in Monte Carlo: Managing Simulations Under Variability and Reliability here, or sign-up for a Solido DAC demo here.
  • Tempusfrom Cadence. Funny story: John Cooley claims to have scooped everyone on this new product announcement even though information about Tempus was sent to us the prior week. Paul McLellan was briefed at Cadence HQ but was asked not to publish until announcement day. Other credible news sources got the embargoed press release days before the announcement as well.
  • Floorplan Compilerfrom Oasys. At the recent EDA Consortium Annual CEO Forecast and Industry Vision event members were asked to vote for the “Hottest EDA Startup”. The vendor receiving the most votes was Oasys Design Systems. I have heard great things about this product from customers and ecosystem partners alike. Request a demo at DAC here.
  • SoSby Cliosoft. This is the only DM platform integrated with all major analog and custom IC design flows and the most talked about tool by customers on SemiWiki. Request a DAC demo here.
  • Timing Explorerfrom ICScape. This is the only timing ECO closure solution that is placement and routing aware and is able to cut the timing closure phase in half, typically within 2-4 iterations. Request a DAC demo here.
  • PathFinderfrom Apache. ESD (electrostatic discharge) is a top trending topic on SemiWiki right now. Request a DAC demo here.
  • VersICfrom Methodics is the first complete verification management system I have seen for analog design. Request a DAC demo here.
  • Anything FinFET,which is probably the most intriguing technology we will see this decade!

    And don’t forget the all important IPL Dinner:

    iPDKs: A Thriving PDK Standard

    Tuesday, June 4, 2013
    6:00 p.m. to 7:30 p.m.
    Austin Hilton Hotel, Grand Ballroom G

    Foundry support for interoperable PDKs (iPDKs) continues to grow. As of 2012, 4 of the top 5 semiconductor pure-play foundries have joined the IPL Alliance, and all the top 5 foundries have provided customers with iPDKs. iPDKs benefit the entire custom design ecosystem. Semiconductor foundries and IDMs create iPDKs to reduce their PDK development, validation, support and distribution costs while enabling advanced design flows and multiple EDA tool support. Chip designers now enjoy access to best-in-class tools, interoperable flows, and improved productivity.

    At the 7th Annual IPL Luncheon, IPL Alliance presenters will highlight the benefits of the iPDK standard and their experiences in developing and deploying foundry iPDKs. The IPL Alliance will also present an update on current and future projects as well as collaboration with other industry initiatives.

    Agenda:

    [TABLE] cellpadding=”2″ cellspacing=”1″ style=”width: 500px”
    |-
    | valign=”top” style=”width: 30%” | 6:00 pm – 6:10 pm
    | valign=”top” style=”width: 70%” | Complimentary dinner
    |-
    | valign=”top” | 6:10 pm – 6:15 pm
    | valign=”top” | Introductions – IPL update and roadmap
    |-
    | valign=”top” style=”width: 30%” | 6:15 pm – 6:30 pm
    | valign=”top” style=”width: 70%” | iPDK advantages – Foundry perspective
    |-
    | valign=”top” | 6:30 pm – 6:45 pm
    | valign=”top” | iPDK Benefits – Customer perspective
    |-
    | valign=”top” style=”width: 30%” | 6:45 pm – 7:00 pm
    | valign=”top” style=”width: 70%” | iPDK and AMS Reference Flows
    |-
    | valign=”top” | 7:00 pm – 7:15 pm
    | valign=”top” | iPDK and OPDK
    |-
    | valign=”top” style=”width: 30%” | 7:15 pm – 7:30 pm
    | valign=”top” style=”width: 70%” | Q&A
    |-

    Register Now!

    Attendance at this event is free, but registration is required. Seating is limited, so reserve your seat early.

    lang: en_US


  • Robust Reliability Verification: Beyond Traditional Tools and Techniques

    Robust Reliability Verification: Beyond Traditional Tools and Techniques
    by SStalnaker on 05-31-2013 at 7:10 pm

    Robust Reliability Verification: Beyond Traditional Tools
    by Matthew Hogan, Mentor Graphics

    At all process nodes, countless hours are diligently expended to ensure that our integrated circuit (IC) designs will function in the way we intended, can be manufactured with satisfactory yields, and are delivered in a timely fashion while meeting the market need. Traditional IC verification relies on a collection of well-known and well-understood tools. Design rule checking (DRC), layout vs. schematic comparison (LVS), electrical rule checking (ERC), parasitic extraction (PEX), design for manufacturing (DFM) and simulation (most often SPICE and timing closure) are all used as part of this cohesive verification flow that provides us the insight required to find and correct any errors or omissions in our design process. Many design errors lead to hard failures in manufacturing, and can be readily identified and fixed, like a metal width that is too small for a process node layer, cells that were incorrectly placed, or shorts across other elements in the design. Finding and fixing these issues is the mainstay of IC verification.

    The legacy of simulation
    SPICE simulation, and the associated parasitic extraction that it uses, plays a vital role in identifying less obvious errors—those that deal primarily with reliability. Ensuring that you have the correct simulation vectors to provide sufficient coverage while validating the waveforms or analyzing messages from your simulation environment can be time-consuming and CPU-intensive activities, where results often require both expert interpretation and the keen eye of someone who understands the subtleties of each particular design.

    Finding scalable alternatives
    Whichever Greek philosopher first said that necessity is the mother of invention must have foreseen the challenges that the IC industry would one day face. Time and again, when faced with a new set of requirements not addressed by existing tools, engineers have leveraged their imaginations to create innovative solutions, designs, and process flows.

    The same is true for reliability verification. With larger designs, smaller process nodes, and the increased pressure on time-to-market schedules and productivity targets, many design teams are turning to new alternatives that provide critical advantages over existing tools:

    • a simple-to-use environment for the designer and verification engineer,
    • fast runtime (that can scale to the full chip),
    • a cohesive platform that is able to validate a wide range of issues

    One tool that has found a strong role in reliability verification is Calibre[SUP]®[/SUP] PERC™. With its ability to evaluate both the logical intent and physical implementation of the design, Calibre PERC provides a unique and powerful reliability verification platform not previously available. While there are many applications where Calibre PERC technology is successfully leveraged, one of the most common uses is the automated identification and resolution of typical reliability design challenges:

    • Electrostatic discharge (ESD)
    • Electrical overstress (EOS) and power intent
    • Voltage-aware DRC

    Many of these topics may be quite familiar to you, or perhaps you already have solutions in place today to help with these issues, but let’s go through them one at a time to provide a broader understanding for all.

    ElectrostaticDischarge
    Designers have always needed to ensure that designs are robust from an ESD perspective. To provide that surety, they need to know what structures the design requires to protect pins from an ESD event, and they need to make sure that the implementation of those structures is correct. They also need to verify that the design complies with the topology rules (that is, the correct combination of protection devices are in place), and that these devices are robust enough to handle the ESD event.

    Topology rules exist to help the designer verify that the layout correctly implements the design intent. For example, do you have robust ESD structures in place (primary and secondary) to protect the pins? Are there anti-parallel (back-to-back) diodes in place for multiple power domain designs? Do you have level shifters in place for signals? Are the metal widths sufficient? Are there enough vias?

    Point-to-point and current density simulations ensure metal lines and vias are sufficiently robust to handle the expected energy, should an ESD event occur. However, these types of issues are difficult to identify using traditional simulation technologies.

    Generalized ESD cells are often designed for this use, but the designer must still ensure that the cell is placed correctly into the design. Additionally, the chip design may change after the ESD IP is placed, requiring the designer to adjust the ESD IP to ensure it fits within the new design parameters (area, performance, etc.). In these situations, it is essential to validate the philosophy of the ESD intent, not just check that the ESD IP has remained intact and unchanged. ESD specialists are often called upon to provide custom solutions for each design, while keeping an eye out for known issues and previous concerns. The best IP in the world can be compromised by a simple implementation oversight.

    While never an ideal solution, many designers have always relied on visual inspection and manual methods to evaluate the accuracy of the implementation of ESD structures. The large number of pin pairs in today’s devices make this solution a daunting, if not impossible, task. The designer simply can’t select a “typical” connection and evaluate just a few; rather, every reasonable combination must be evaluated. This challenge was one of the catalysts that led to a rethinking of how ESD structures are evaluated.[1]

    Calibre PERC can automatically select and analyze all of the required combinations. For schematic checking, the rules are directed more towards verifying the presence of the appropriate protection schemes from a topological perspective. Users can perform checks on circuitry directly connected to pads, as well as checks on the ESD network. For layout checking, the rules focus on verifying the point-to-point parasitic resistance between the pad and the ESD device, checking current density between pad and the ESD device, detecting pmos/nmos devices sharing the same well, detecting pmos/nmos field oxide parasitics, detecting latch-up issues, and more.

    EOS and powerintent
    Designs that incorporate multiple power domain checks are particularly susceptible to subtle design errors. Often, these subtle errors don’t result in immediate part failure, but performance degradation over time. Effects such as Negative Bias Temperature Instability (NBTI) can lead to the threshold voltage of the PMOS transistors increasing over time, resulting in reduced switching speeds for logic gates, and Hot Carrier Injection (HCI), which alters the threshold voltage of NMOS devices over time. Soft breakdown (SBD) also contributes as a time-dependent failure mechanism, contributing to the degradation effects of gate oxide breakdown.

    Transistor-level power intent verification is a critical need, especially in designs that make extensive use of IP. The IP must be hooked up correctly within the design. Thin oxide gates and high power applications require tight controls for voltage and power domains. Many of these issues are difficult to identify in the simulation space or with traditional PV techniques.

    Power-aware checking requires the ability to use the design’s netlist to recognize specific circuit topologies, such as level shifters, I/O drivers, and other structures, and then relate those to the corresponding GDS geometries that make up the layout, to verify that those specific elements have been included and have been implemented correctly. Unlike the foundry DRC decks, the definition of these checks do not all come from the foundry, but must be tailored to the specific design styles and practices of the designer’s company, so any tool performing this function must be highly flexible and easily programmable. A transistor-level power-aware checking tool must also be able to statically propagate voltage values from the various supplies to every node in the circuit to facilitate a variety of EOS checks.

    For example, one common problem for designers trying to debug power violations at the transistor level with simplistic tools is a lack of knowledge of the intention (the functionality) of the circuit where the violation is found. Simply checking for transistors connected to multiple domains results in a large number of false errors at the boundary between domains, where level shifter structures intentionally include transistors exposed to both low and high voltages. A power-aware checking tool like Calibre PERC can prevent such false errors by using an automated circuit recognition technique to identify particular topologies. The circuit recognition functionality within Calibre PERC uses the SPICE syntax as an easy way to define complex circuit structures. Whenever power violations are detected in enable-LS, NAND or NOR structures, false errors can be quickly waived using topological recognition.

    Additionally, the unified power format (UPF) provides a way to annotate a design with power intent that is independent of any hardware description language (HDL). It is typically used at all levels of the design flow. A UPF specification at the register transfer logic (RTL) level defines the power architecture of a given design, and drives synthesis and place-and-route to achieve correct implementation. In automated reliability verification, using the same UPF specification for transistor level physical verification ensures the original power intent is preserved with the final implementation.

    UPF specifications can be leveraged as an integral part of Calibre PERC’s understanding of power intent. Along with the design layout data and verification rule deck, Calibre PERC examines the UPF definitions of supply networks (consisting of power switches, supply ports, and supply nets) and checks each supply port’s supply states and its connected supply net. Most importantly, it analyzes the power state tables defined in terms of these states to ensure it captures the legal combinations of supply voltages in the entire design. With integrated support for UPF, Calibre PERC can automatically assign voltages based on a design’s power intent, greatly improving verification coverage and robustness.

    Voltage-aware DRC
    For smaller process nodes and high reliability designs, the spacing requirements between nets vary as the nets traverse through the design. The required spacings are dependent on the operating voltage ranges, and devices operating at different voltages must be properly protected. For example, many designs have high voltage areas, such as flash memories, that are particularly susceptible. Designers must identify vulnerable nets and devices, and perform the appropriate spacing and guarding checks on the layout. With traditional verification methods, this means creating physical layout markers to perform voltage-aware DRC.

    Using its novel circuit topology-aware voltage propagation capability, Calibre PERC can automatically perform voltage analysis and apply the results against the schematic or extracted layout netlist. Target nets and devices for the voltage-aware DRC checks are selected from the layout through the direct integration of netlist-based voltage analysis, using either vectored or vectorless static voltage propagation. The voltage-aware DRC rules are then applied to the selected layout objects. Such analysis and verification is used to identify areas of the design at risk for time dependent dielectric breakdown (TDDB).

    Summary
    Robust reliability verification is available today as a comprehensive solution. Verification tools such as Calibre PERC include specific technologies to make fast, automated reliability verification practical. With a reliability verification methodology comprising a single tool with a unified rule deck and integrated debug environment, Calibre PERC helps designers find subtle design optimization opportunities without SPICE circuit simulation, while also enabling them to achieve the accurate and comprehensive verification necessary to ensure a repeatable and reliable design.

    References

    [LIST=1]

  • Muhammad, M.; Gauthier, R.; Junjun Li; Ginawi, A.; Montstream, J.; Mitra, S.; Chatty, K.; Joshi, A.; Henderson, K.; Palmer, N.; Hulse, B., “An ESD design automation framework and tool flow for nano-scale CMOS technologies,” Electrical Overstress/ Electrostatic Discharge Symposium (EOS/ESD), 2010 32nd , vol., no., pp.1-6, 3-8 Oct. 2010
    URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5623716

    Going to DAC and interested in learning more about Calibre PERC? Matt will be presenting on reliability verification several times at the Mentor Graphics booth (#2046):

    • Monday, 2:00-3:00 (Comprehensive Circuit Reliability with Calibre PERC)
    • Tuesday, 2:00-3:00 (Advanced Circuit Reliability with TowerJazz; complements lunch seminar)
    • Wednesday, 10:00-11:00 (Comprehensive Circuit Reliability with Calibre PERC)

    Registration is required for booth presentations: sign up here

    Matt will also be participating in a panel on Monday from 3:00-4:00 in the front of the Mentor booth:
    Achieving IC Reliability in High Growth Markets

    No registration is required for panels – simply show up, listen, and learn! The panel will be followed immediately by Mentor’s complimentary Happy Hour, where you’ll have a chance to enjoy an adult beverage, and engage with Matt and other Mentor experts in a relaxed environment.


    Matthew Hogan is a Product Marketing Manager for Calibre Design Solutions at Mentor Graphics, with over 15 years of design and field experience. He is actively working with customers who have an interest in Calibre PERC. Matthew is an active member of the ESD Association—involved with the EDA working group, the Symposium technical program committee, and the IEW management committee. Matthew is also a Senior Member of IEEE, and a member of ACM. He holds a B. Eng. from the Royal Melbourne Institute of Technology, and an MBA from Marylhurst University. Matthew can be reached at matthew_hogan@mentor.com


  • Cooley’s Cheesy Must See List for DAC is Out

    Cooley’s Cheesy Must See List for DAC is Out
    by Paul McLellan on 05-31-2013 at 6:23 pm

    One of the other increasingly successful channels (besides Semiwiki of course) for EDA, IP and semiconductor companies to reach potential customers is John Cooley’s DeepChip. Every year he puts a lot of effort into trying to find out who is exhibiting what at DAC and which stuff seems like it is new and maybe important, and he produces a long guide (a couple of dozen pages). He even lists by name the most appropriate contact person at each company for the product that he is talking about. Sure, they are mostly marketing guys, but they are the marketing guys who specialize in that product, not just generic contacts at the company.

    Print it out and read it on the plane to Austin and you’ll have a much better idea of what stuff is worth your time. Of course it is his opinion yours may differ. But I’ll bet you’ve not put as much thought into the entire spectrum of what is being show at DAC

    This years guide is now out just a couple of days before DAC.

    Gary Smith’s what to see @ DAC 2013 List is here.

    SemiWiki Top Ten Must See @ #50DAC List is here.

    I have no idea if this is accurate, but apparently word is Wally Rhines, Lip-bu Tan, Kathryn Kranen, JL Grey, Gary Smith, Joe Costello, Suk Lee, Richard Goering, Raul Camposano, Dean Drako all donated to charity to be at Jim Hogan’s Hot Zone VIP party at 8:00 to 1:00 on DAC Monday at Austin City Limits. I donated too so go ahead and donate yourself and I’ll see you there.



    The Semiconductor Wiki Project
    , the premier semiconductor collaboration site, is a growing online community of professionals involved with the semiconductor design and manufacturing ecosystem. Since going online January 1st, 2011 more than 600,000 unique visitors have been recorded at www.SemiWiki.com viewing more than 5M pages of blogs, wikis, and forum posts.

    Gary Smith EDA (GSEDA) is the leading provider of market intelligence and advisory services for the global Electronic Design Automation (EDA), Electronic System Level (ESL) design, and related technology markets.

    DeepChip.com is a 20 year old clearinghouse where semiconductor chip designers contribute data-intensive papers and articles of first-hand evaluations and production benchmarks of commercial EDA tools.


    ARM Partners with Carbon on Cortex-A57

    ARM Partners with Carbon on Cortex-A57
    by Paul McLellan on 05-31-2013 at 3:37 pm

    Just in time for DAC, Carbon have announced that they have expanded their partnership with ARM to create and deliver models for the ARM Cortex-A57 processor and related IP. One piece of related IP is the Cortex-A53 which can be configured in big.LITTLE multi-core setups to achieve the sweet spot of higher performance and lower power. A57 when you need it, A53 when you don’t.

    But ARM’s IP family has got quite extensive and the agreement also includes the Corelink CCN-504 cache-coherent network and the Mali-T628 GPU. Carbon will take all this technology and compile 100% accurate virtual models as well as Carbon Performance Analysis Kits (CPAKs) that can boot Android or Linux in seconds while still preserving Carbon’s secret sauce, the capability to switch to 100% accurate representation at any breakpoint and then proceed with almost any level of detail of the design exposed. A bit like big.LITTLE: accuracy when you need it, high speed when you don’t.

    A57 and A53 are ARM’s first 64-bit cores. They can run 32-bit legacy applications too. In fact 32-bit versions of the core are also available. Who knows when we will really need 64-bit in our phones, but in the meantime the focus is on producing very low power servers for specialist datacenter applications. The ARM licensees focused on this market all have value propositions something like 10% cost, 10% power and 10% of the physical volume of equivalent traditional (Intel) solutions. For internet applications, being able to handle millions of transactions simultaneously at low power and cost is more important than the single thread performance of any one transaction (weather forecasting or simulating nuclear bombs is the other way around, but the fast growing part of the market is datacenters for Apple, Facebook, Amazon, eBay etc) making for a real opportunity.

    A system including A57, A53 and Mali is pretty complicated since everything has cache coherent interoperability. There may be more than one core of A57 or A53 of course (as in the latest Samsung phone that contains 4 high performance and 4 low power cores although they are earlier Cortex-A15/A9 not the A57/53).

    Carbon makes it feasible to run full software loads on the design, even very early when decisions are being made about which cores to use and how many. They can then be used for early (pre-silicon, in fact pre-RTL, pre-everything except basic architectural decisions) software development and high visibility post-silicon debug. Plus, they can drop into fully accurate mode to debug subtle problems with hardware or device drivers or to get accurate timing for critical algorithms.

    As usual, models will be available from Carbon’s IP Exchange web portal which is here. Models of the A57 and Mali-T628 are available today for select early access partners.

    Carbon will be at DAC but they won’t have their own booth. They will be in the…surprise…ARM booth, #921.


    10 years, 100,000 miles, or <1 DPM

    10 years, 100,000 miles, or <1 DPM
    by Don Dingee on 05-30-2013 at 10:00 pm

    Auto makers have historically been accused of things like planned obsolescence – redesigning parts to make repairs painfully or even prohibitively expensive – and the “warranty time-bomb”, where major systems seem to fail about a week after the warranty expires. Optimists would chalk both those up to relentless innovation, prudent engineering, and cost containment.

    Continue reading “10 years, 100,000 miles, or <1 DPM”


    SEMulator3D – A Virtual Fab Platform

    SEMulator3D – A Virtual Fab Platform
    by Pawan Fangaria on 05-30-2013 at 8:30 pm

    Yes, it’s a pleasant surprise; it is Virtual Fabrication Platform, one of the new innovations in 2013. I was looking around for what kind of breakthrough technologies will be announced in DAC this year. And here I came across this new kind of innovative tool which can produce final virtual fabricated 3D structures after following all the complex steps of actual fabrication process based on process parameters and design data. Amazing, isn’t it?

    Coventor has introduced SEMulator3D (currently available for shipping) at the right time when we are talking about 3D transistors (Tri-Gate, FinFET and the like), High-k/Metal Gate and sub 22nm processes which come with their own challenges of fabrication. While complexities of process technology are hitting the limits, in order to keep the Moore’s law (according to which the number of transistors on an IC will double in every two years) alive, we must reduce the number of iterations and shorten the time between design and final physical silicon.

    It’s a commendable job by Coventor, who, amid increasing complexities at sub 22nm process nodes and 3D Gates, has come up with this new process modelling paradigm. It handles the complexities of integrated 3D front-end-of-line (FEOL) manufacturing processes quite well. On the virtual automated platform, it reduces the cycle time between fabless design and foundry from months to days or hours, hence seamlessly increasing collaboration between the two teams and reducing cost dramatically.

    The SEMulator3D engine employs advanced physics-driven predictive modelling techniques, such as voxel and surface evolution, which bring high order of physical accuracy. Voxel modelling is a fast and robust digital approach, capable of scaling to the requirements of integrated processes and large silicon areas. Surface evolution is a more analog approach, capable of modelling a wide range of physical process behaviours. I would like to write more about these technologies in future, but for now I can foresee that the concept of SEMulator3D will gain importance and proliferate widely to meet the challenges of today’s semiconductor design and fabrication. SEMulator3D also provides automatic process variation analysis with parallel modelling and virtual metrology that enable in-line, local measurement of critical dimensions, mimicking real metrology operations.

    SEMulator3D is a must to reduce silicon learning cycles, faster time-to-market (especially for new 20nm and below process nodes) and saving $$ spent in reaching manufacturing readiness.

    For more information –
    Visit Coventor booth # 1326 at DAC 2013

    Attend a technical presentation by CTO of Coventor, Dr. David Fried on “Virtual Fabrication: Integrated Process Modelling for Advanced Technology” at SEMICON West – San Francisco, CA, July 9 – 12

    See Coventor press release here.