chip design magazine april-may 2010
TRANSCRIPT
April/May 2010
INTEGRATED IP GOES VERTICAL
PROTOTYPING OPTIONS FOR HARDWARE/SOFTWARE DEVELOPMENT
MAKING ABSTRACTION PRACTICAL
AVOID THAT EMBARRASSING CALL TO THE FIRMWARE VENDOR
Also in this issue:
Lead Story:
www.chipdesignmag.com
Affiliate Sponsors:
chipdesignmag.com: community portals, blogs, videos, trends-surveys, e-letters (IP, PLDs, and Chip Designer), resource catalog, back issues, ...
DESIGNMEETS
AUTOMATION
SEE INSERT FOR DETAILS ON THE MAIN EVENT FOR ELECTRONIC
DESIGN
In technicalcooperation with:
Sponsored by:
www.dac.comFOR MORE DETAILS, VISIT:
KEYNOTESTUESDAY, JUNE 15
WEDNESDAY, JUNE 16
THURSDAY, JUNE 17
Exciting Events on the Exhibit FloorMONDAY, JUNE 14-169:00AM - 6:00PM
Technical Program Highlights
ECHOES OF DAC’S PAST: FROM PREDICTION TO REALIZATION, AND WATTS NEXT?
Bernard Meyerson
FROM CONTRACT TO COLLABORATION: DELIVERING A NEW APPROACH TO FOUNDRY
Douglas Grose
DESIGNING THE MOTOROLA DROID
Iqbal Arshad
ANAHEIM CONVENTION CENTER
JUNE 13-18Anaheim, CA USA
PanelsTuesday, June 15
Wednesday, June 16
Thursday, June 17
Embedded/SOC Enablement Day - Thursday, June 17“Embedded Systems Meet Hardware”
TutorialsMonday, June 14
Friday June 18
User Track SessionsPosters and presentations by and for users of EDA tools.
Tuesday, June 15
Wednesday, June 16
Thursday, June 17
Management DayTuesday, June 15
2 • April / May 2010 Chip Design • www.chipdesignmag.com
www.chipdesignmag.com
Publisher & Sales DirectorKaren Popp (415) 255-0390 x19
EDITORIAL STAFF
Editor-in-ChiefJohn Blyler (503) 614-1082
Consulting EditorEd Sperling
Managing EditorJim Kobylecky
Coordinating Regional EditorPallab Chatterjee
Associate Editor—China Jane Lin-Li
Executive Editor – iDesignClive "Max" Maxfield
Contributing EditorsCheryl Ajluni, Dave Bursky, Brian Fuller, Ann Steffora
Mutschler, Craig Szydlowski
Editorial BoardTom Anderson, Product Marketing Director, Cadence • Cheryl
Ajluni, Technical Consultant, Custom Media Solutions • Karen Bartleson, Standards Program Manager, Synopsys • Chuck Byers, Director Communications, TSMC • Lisa
Hebert, PR Manager, Agilent • Kathryn Kranen, CEO, Jasper Design Automation • Tom Moxon, Consultant, Moxon Design
• Walter Ng, Senior Director, Design Services, Chartered Semiconductor • Scott Sandler, CEO, Novas Software • Steve
Schulz, President, SI2 • Adam Traidman, Chip Estimate
CREATIVE/PRODUCTION
Graphic DesignersKeith Kelly & Brandon Solem
Production CoordinatorSpryte Heithecker
SALES STAFF
Advertising and ReprintsKaren Popp (415) 255-0390 x19
Audience Development / CirculationJenna Johnson • [email protected]
President Vince Ridley (415) 255-0390 x18
Vice President, Marketing & Product DevelopmentKaren Murray • [email protected]
Vice President, Business DevelopmentMelissa Sterling • [email protected]
Vice President, Sales Embedded Systems Media Group
Clair Bright • [email protected]
TO SUBSCRIBE OR UPDATE YOUR PROFILEwww.chipdesignmag.com/subscribe
SPECIAL THANKS TO OUR SPONSORS Chip Design is sent free to design engineers and engineering managers in
the U.S. and Canada developing advanced semiconductor designs. Price for international subscriptions is $125, US is $95 and Canada is $105.
Chip Design is published bimonthly by Extension Media LLC, 1786 18th Street, San Francisco, CA 94107. Copyright © 2009 by Extension Media
LLC. All rights reserved. Printed in the U.S.
IN THIS ISSUE
14
Cover Story—Focus Report
Integrated IP Goes Vertical
By Ed Sperling Departments
4 Chip Design Online
6 Editor's Note: EDA Tool Vendor
- A Rose by any other Name? By John Blyler, Editor in Chief
10 In the News--People in the News
By Jim Kobylecky
12 SoCs Move Beyond Digital and
Memory Blocks
By John Blyler
33 Dot.Org-- Creatively Supporting
the EDA Community By John Darringer, President of the
IEEE Council on EDA
35 Top view-- NoC Technology
Offers Smaller, Faster, and More
Efficient Solutions
By K. Charles Janac, Arteris Holdings
36 No Respins— MEMS Is Poised
to Cross the Chasm By Dr. Joost van Kuijk, Coventor
23
Features
Prototyping Options for Hardware/Software Development
How to choose the right prototype for pre-silicon software development.
By Frank Schirrmeister, Synopsys
Making Abstraction Practical
A Vision for a TLM-to-RTL Flow
By Lauro Rizzatti, EVE-USA
Avoid That Embarrassing Call to the Firmware Vendor (and Other Tricks of Low-Power Verification) Simulation-based hardware/software co-verification lets you address power problems early and well.
By Marc Bryan and Barry Pangrle, Mentor Graphics
Low-Power Tradeoffs Start at the
System Level To succeed at the end, your design methodology must weigh all design constraints from the very beginning.
By Ed Steve Svoboda, Cadence Design Systems
Take A New Approach to the Power-Optimization of Algorithms and Functions
By Rishiyur S. Nikhil, Bluespec, Inc.
19
16
26
READ ONLINEwww.ch ipdes ignmag.com
29
Six powerful Virtex®-6 FPGAs, up to 24 Million ASIC gates, clock speeds to710 Mhz: this new board races ahead of last generation solutions. The Dini Group
has implemented new Xilinx V6 technology in an easy to use PCIe hosted or standalone board that features:
• 4 DDR3 SODIMMs, up to 4GB per socket
• Hosted in a 4-lane PCIe, Gen 1 slot
• 4 Serial-ATA ports for high speed data transfer
• Easy configuration via PCIe, USB, or GbE
• Three independent low-skew global clock networks
The higher gate count FPGAs, with 700 MHz LVDS chip to chip interconnects, provideeasier logic partitioning. The on-board Marvell Dual Sheeva processor provides multiplehigh speed interfaces optimized for data throughput. Both CPUs are capable of2 GFLOPS and can be dedicated to customer applications.
Order this board stuffed with 6 SX475Ts—that’s 12,096 multipliers and more than 21million ASIC logic gates—an ideal platform for your DSP based algorithmic accelerationand HPC applications.
Don’t spin your wheels with last year’s FPGAs, call Dini Group today and run yourbigger designs even faster.
www.dinigroup.com • 7469 Draper Avenue • La Jolla, CA 92037 • (858) 454-3419 • e-mail: [email protected]
DNV6F6PCIe
4 • April / May 2010 Chip Design • www.chipdesignmag.com
CHIP DESIGN ONLINEwww.chipdesignmag.com
Blogs: www.chipdesignmag.com/blogs
JB’S CIRCUIT
John finds blossoms and thorns
woven into today’s EDA.
COLLABORATIVE ADVANTAGE
Steve Schulz explores the
boundaries work against
cooperation in EDA standards.
NPD MANAGEMENT CORNER
Jeff Jorvig takes us inside new
product development.
EDA THOUGHTS
Daniel Payne takes Single Event
Upset for a Prius testdrive.
WIZARDS OF MICROWAVE
Marc Petersen and Colin
Warwick play Flip-Chip.
PALLAB’S PLACE
Pallab Chatterjee finds a
welcome focus on innovative
customer solutions.
TUNING IN TO JIM
Jim Lipman sees a positive
direction for semiconductor
development.
Look for Women in Electronic Design,
a new blog series that will be featuring
many of the engineers and companies
who are the future of EDA.
KOBY’S KAOS
In defense of arrogance.
VISIT
www.chipdesignmag.com
TODAY
PORTALSVisit our growing technology
communities with other chip
architects, engineers and managers.
System-Level Design portal:
www.chipdesignmag.com/sld
Low-Power Design portal:
www.chipdesignmag.com/lpd
Be sure to visit our growing selection
of community blogs including Editor’s
Note by Ed Sperling, ESL Edge by Jon
McDonald, A View From the Top by Frank
Schirrmeister, The Vipster by Vipin Tiwari,
VC Corner with Jim Hogan and Peter L.
Levin, Voltedge by Bhanu Kapoor, Power
Play with Arvind Narayanan, Embedded
Logic by Markus Levy, and Absolute Power
with Cary Chin and Darin Hauer.
BUT WAIT, THERE’S MORE
New and familiar bloggers are coming to
the individual resource pages EECatalog at
http://eecatalog.com/
ASIC-ASSP Prototyping Survey Report
Report ID: CS051608
Price: $299
EDA Tools-Trends Survey
Report ID: CDT063009
Price: $299
IDESIGN: www.chipdesignmag.com/idesign
Power Architecture ISA 2.06 Stride N
Prefetch Engines to Boost Application's
Performance
Three top engineers of the IBM India
System and Technology Labs take us inside
superscalar POWER design.
Billions of Cycles for Billions of Gates
How does a design team execute one-
billion cycles on a one-billion gate design?
Discover what happens when hardware-
assisted verification comes into play.
e-Newsletterswww.chipdesignmag.com/enewsletters
Track the latest industry news and views with our e-newsletters.
CHIP DESIGNER
• The 5-minute Guide to Chip-level
Audio Test
• Thanks for the Memories, But…
• IC Design Flow Must Evolve For
Challenges of 28nm and Below
IP DESIGNER & INTEGRATOR
• Designers – Start Your
Characterization Engines
• SoC Designers Must Have Tangible
Quality Metrics for IP
PROGRAMMABLE LOGIC DEVICE DESIGNER
• Today’s FPGAs Offer High-
Performance DSP Capabilities
• The Academy Award Goes to …
Verification Engineers of all Stripes
CHIP DESIGNER PLUS
• Still Room for Startups?
• Typing words with your brain
• How to Stay Current When You’re
Out of Work
• The Latest in Wearable Media
Technology
Chip Design • www.chipdesignmag.com April / May 2010 • 5
PORTALS
www.SLDCommunity.com
www.LPDCommunity.com
Participate in these Growing Online Communities for System-Level
and Low-Power Designers
6 • April / May 2010 Chip Design • www.chipdesignmag.com
By John Blyler, Editor-in-ChiefEDITOR'S NOTE
What is happening in the EDA
industry? Irmgard Lafrentz, President
and Founder of Globalpress, poised
this question to me in a recent phone
call. She did a good job of capturing the
essence of the conversation in a recent
blog: Something’s happening in EDA,
and this time it’s good! (http://globalink.globalpresspr.
com/blog/2010/04/somethings-
happening-in-eda-and-this-
time-its-good.html) I want
to explore this question in a
bit more detail.
First, let’s consider the
media side of this question.
With the collapse--but
not total annihilation--of
the print business model as a
primary means for funding the development of meaningful
content, EDA companies are finding fewer and fewer
venues for their technology and product announcements.
Couple that challenge with the necessity of reaching a
global audience with their message. This means that EDA
companies must look for coverage beyond the traditional
sources--hence the push into online and social media
outlets. (Yes – there are other reasons for the push, too.)
Secondly, many EDA companies are making a serious
push into vertical markets with their technology and
products, markets like medical, industrial, automotive and
communications.
This “shift” way from being seen as an “EDA tool vendor”
to instead being perceived as a “system solution provider” is
both evolutionary and essential for business survival. EDA
companies can no longer focus solely on the design and
manufacture of the chips. Instead, they must consider the
chips in relationship to the package and board--and even
in terms of both hardware and software. This is one reason
why IP has become so critical in the EDA tool chain.
This move away from the nomenclature of the “EDA tool
vendor” can be seen in the restructuring of most of the big
technical trade shows, too--like DAC and ESC.
So, if you don’t want to be known as an “EDA tool vendor,”
then what is the correct phrase? System Solution Provider?
That sounds rather PR-ish. But what phrase will capture
the essences of today’s EDA company? I’m not sure. What
do you think?
++++++++++++++
Selected responses:
None of the EDA vendors
today will be able to become
a solutions provider for any
domain any time soon.
While they have a large portfolio of products, this portfolio
is absolutely incomplete/ has gaping holes when looked at
EDA vendor by EDA vendor.Any one vendor who wants to
succeed really as a solutions provider will have to address
the COT aspect and really cooperate with other EDA
vendors to jointly integrate different vendor products into
one solution. Something CAD teams do today, and do well.
They are currently the real EDA solution providers in the
IC industry.
Will such cooperation happen? With the current back-
stabbing/ throat-cutting attitude in the EDA industry, I
really wonder.
Is this just a problem of scale, i.e. is the EDA industry
too small?There are plenty of other examples in the
SW domain where collaborations/ networks/ seamless
integration between multiple vendors is reality.
Once EDA realizes this as well, they have a chance to
become solution providers. -- T
EDA Tool Vendor – A Rose by any other Name?
+
S
N
t
a
d
EDA companies can no longer focus solely on the design and manufacture of
the chips.
Let Our Eyes Catch Yours
PHY SerDes TX/RX PLL/DLL Analog Blocks
4423 Fortran Court, Suite 170San Jose, CA 95134 • 408.942.9300
www.Mixel.com
Mobile • MIPI • MIPI/MDDI Unifi ed Solution • MDDI
Timing • Frequency Synthesis • Fractional-N • Spread-spectrum • De-skewing • CDR
Storage & Networking • PCI-Express • DisplayPort • XAUI • Fibre-Channel • SATA • GPON
Interface • LVDS • CE-ATA • CardBus • PATA
• DDR2 • SSTL • HSTL • PCI-X
™
Come visit us at DAC Booth # 416
In technicalcooperation with:
Sponsored by:
GENERAL CHAIR’S WELCOMELET’S MEET AT DAC!
Dear Colleague:The 47th edition of the Design Automation Conference in Anaheim is just around the corner, and I look forward to welcoming you there. As the central “meeting place” for electronic design and design automation where the industry puts on its grand annual show, there are many facets to DAC. It’s the place new contacts are made, where deals are sealed, where theory meets practice, where colleagues across the industry network, where the seeds of great new ideas are sowed – and much more. DAC is our annual signpost that points the way to the future.
As organizers of the event, we work with DAC’s sponsors and hundreds of volunteers to make it worth your time to attend. This year, in addition to reinforcing traditional strengths, we have added a number of exciting new elements. Here’s a sample of what you can see at DAC:
• The keynote lineup features three distinguished and accomplished industry luminaries: Doug Grose, CEO of GLOBALFOUNDRIES will address the central role of the foundry in electronic design on Tuesday. Bernie Meyerson, Vice President for Innovation at IBM Corporation, will discuss his vision for next-generation IT infrastructure for EDA and the move towards cloud computing, and Iqbal Arshad, Corporate Vice President of Innovation Products at Motorola, will overview his experiences in driving the Motorola Droid from concept to product.
• A vibrant exhibition showcases nearly 200 companies, including all of the largest EDA vendors and a significant foundry presence. The Exhibitor Forum theater features focused technical presentations from exhibitors, while IC Design Central’s exhibit area and a presentation stage that brings together the entire ecosystem for SOC enablement, including IP providers, design services provides, and foundries.
• A special Embedded/SOC Enablement Day on Thursday is designed to further advance DAC’s partner eco-system, attracts a mix of chip creators, ecosystem suppliers, and research-focused participants.
• A robust and exciting technical program includes an exciting array of panels and special sessions that complement a carefully selected subset of the contributed research papers.
• The User Track program, specifically designed by and for EDA tool users, features presentations and poster sessions that highlight outstanding solutions to critical design and methodology challenges, and case studies of innovative tool use. In its second year, it is 50% larger than last year’s acclaimed program.
• An excellent slate of tutorials covers topics such as low-power design, ESL, and software development for the EDA professional.
• Management Day includes invited presentations and networking opportunities for decision-makers in the industry, and highlights issues at the intersection of business and technology.
• An impressive constellation of fourteen colocated events and six DAC workshops complements the DAC program: this includes established conferences and symposia such as AHS, DFM&Y, DSNOC, HOST, HLDVT, NANOARCH, SASP, and SLIP, as well as meetings on emerging topics such as bio-design automation, mobile/cloud computing, and smart grids.
As you can see, there’s tons of good stuff in store – come join us in Anaheim!
Sachin S. SapatnekarGeneral Chair, 47th DAC
TECHNICAL PROGRAM HIGHLIGHTSThe technical program for DAC 2010 exceptional-quality technical papers, panels, special sessions, WACI (Wild and Crazy Ideas), full day tutorials and User Track. The program is tailored for researchers and developers in the electronic design and design automation industry, design engineers, and management. It highlights the advancements and emerging trends in the design of electronic circuits and systems.
The core of the technical program consists of 148 peer-reviewed papers selected from 607 submissions (a 24% acceptance ratio). Organized in 35 technical sessions, these papers cover a broad set of topics ranging from system-level design, low-power design, physical design and manufacturing, embedded systems, logic and high level synthesis, simulation, verification, test and emerging technologies.
Popular submission themes included:
1. Power Analysis and Low-Power Design
(83 submissions, 5 sessions)
2. Physical Design and Manufacturability
(72 submissions, 4 sessions)
3. System-Level Design and Analysis
(69 submissions, 4 sessions)
Some of the novel ideas presented in these papers includes cutting-edge research in property checking, global routing, variation characterization, silicon mismatch, cache design for routers, rewiring, logic optimization with don’t cares, Boolean matching, and low energy processor design. The papers reflect the increasing importance of system-level design, low-power design and analysis, and physical design and manufacturability.
KEYNOTES
FROM CONTRACT TO COLLABORATION: DELIVERING A NEW APPROACH TO FOUNDRYDouglas Grose, Chief Executive Officer, GLOBALFOUNDRIES, Sunnyvale, CA
The list of challenges facing the semiconductor industry is daunting. Chip design continues to increase in complexity, driven by product requirements that demand exponentially more performance, functionality and power efficiency, integrated into a smaller area. In parallel, manufacturing technology is facing increased challenges in materials, cost and shorter product lifecycles. This confluence of factors puts the industry at a crossroads and the foundry industry at center stage.
Chip design companies need to redefine relationships with their manufacturing partners, and foundries must create a new model that brings manufacturing and design into an integrated and collaborative process. This presentation will explore the challenges of bringing the next generation of chip innovation to market through leveraging an integrated global ecosystem of talent and technology. The world’s top design companies want more than a contract manufacturer; they want a level of collaboration and flexibility supported by a robust partner ecosystem of leading providers in the EDA, IP and design services sectors.
TUESDAY JUNE 15
ECHOES OF DAC’S PAST: FROM PREDICTION TO REALIZATION, AND WATTS NEXT?Bernard S. Meyerson, IBM Fellow, Vice President-Innovation, IBM Corp., Yorktown Hts., NY
Over the last five years the semiconductor industry has acknowledged, but struggled to deal with, the end of classical device scaling in silicon technology. This has had ramifications across all aspects of the technology spectrum, as a steady stream of innovations, ever more fundamental, have been required to drive accustomed generational improvements in Information Technology (IT). Adding to this challenge on the demand side there has been an accelerating and seemingly insatiable need for IT resources, driven by the emergence of the ‘Internet of Things’. With such heavy and growing IT demands, key metrics such as system power, cost/performance, and application specific benchmarks have become a core focus of emerging solutions. It is these same metrics and constraints that also require advances in the efficiency and optimization of IT. In this talk, I will review how our industry is dealing with each of these challenges, and explore emerging compute paradigms, such as Cloud Computing, that are impacting EDA directly.
WEDNESDAY JUNE 16
DESIGNING THE MOTOROLA DROIDIqbal Arshad, Corporate VP, Innovation Products, Motorola Mobile Devices, Inc., Schaumburg, IL
As mobile internet usage skyrockets and more sophisticated mobile applications are being developed, the device formerly known as the cell phone is at a major technological inflection point. To meet this challenge, we must design devices and services that enable a transformation in the way we work, socially interact, use the web and utilize computing power. A key ingredient to making this happen is the synthesis of new hardware that is tightly coupled with a new software experience or business opportunity. Similarly, when launching new high-technology products, the success of the product largely depends on how well the target consumer is educated about the availability and capability of the new device. This talk will discuss how designing the Droid helped Motorola to address this shift in the market place. THURSDAY
JUNE 17
USER TRACKThe DAC User Track brings together IC designers from across the globe. The User Track program offers a unique opportunity to pick up the latest tips and tricks from the industry expert IC designers, and features over 110 presentations on a wide variety of topics. Designers from Intel, IBM, Samsung, TI, Toshiba, Qualcomm, AMD, Freescale and other leading IC companies will present their experiences on building effective design flows, design methods, and tool usage. Come to DAC to attend the User Track - there is no other way to improve your ‘design IQ’ in just a short amount of time!
The list of topics is longer and more diverse than ever, and there is something in the User Track for every designer. Low-power design is one of the core topics addressed at the system-level, RTL, and during the place-and-route stages of design. The User Track features talks on efficient design for low-power and on power delivery on the chip and through the package. In other talks, designers will address dealing with the significant variability at 32nm and below. You will find interesting presentations on variation-robust design methods, with ways to quickly converge on high-yield designs. Timing closure is another theme that is addressed by speakers from a several perspectives, both front-end and back-end. Designers will present innovative ways for partitioning, budgeting and retiming. Also featured are presentations on several timing-driven ECO physical optimization methods, and system-level case studies that use formal verification and much more. Please check out the program in the new dac.com website for the full details on the presentations.
The Design Automation Conference and the User Track bring together thousands of like-minded professionals, making this event an opportunity you cannot miss. The User Track runs for three packed days as a parallel track within the DAC technical program. Learn from expert designers in person, and find out the truth about design tools. Stroll through the DAC trade show, attend keynotes and cutting-edge technical sessions, or just talk to colleagues from other companies. Whether it’s for the full three days or just a single day, DAC has it all. And since DAC 2010 is held right next to Disneyland, this is a great opportunity to bring your family along.
USER TRACK SESSIONS
• Timing is Everything
• Front-End Design Experiences
• Taming Back-End Verification and DFM
• Case Studies in Formal Verification
• Cornered: Dealing with Variability
• Front-End Testing and Verification
• Power Delivery from Package to Chip
• Advances in System-Level Design and Synthesis
• User Track Poster Sessions
MANAGEMENT DAY Tuesday, June 15
Management Day 2010 is focused on issues at the intersection of business and technology, and is specifically directed to managers and decision-makers. Three sessions make up this year’s event. Two sessions will feature managers representing IDMs, fablight ASIC providers, and fabless companies, as well as senior managers designing today’s most complex nanometer chips, and will discuss the latest solutions and their economic impact. The third session will be a panel that involves the presenters and the audience in a brainstorming discussion.
EMBEDDED/SOC ENABLEMENT DAY Thursday, June 17
The Embedded/SoC Enablement Day is dedicated to bringing industry stakeholders together in one room to shed light on where SoC design is headed. The event comprises presentations from leading SoC enabling sectors including embedded processors, embedded systems, EDA, FPGA, IP, foundry, and design services. Presenters will focus on the optimization of embedded and application-domain-specific operating systems, system architectures for future SoCs, application-specific architectures based on embedded processors, and technical/business decision making processes by program developers. This program consists of three sessions, and provides an opportunity to foster discussions that address all aspects of the SoC development ecosystem.
TUTORIALSThe DAC program includes seven tutorials on timely subjects, including four design topics: 3-D integrated circuits, analog mixed-signal design, system-level design, and low-power design. This year also features tutorials on two special topics. The first is an overview of software engineering that includes introductions to agile, lean, scrum, and other software best practices—topics that will offer immediate and practical value to students, EDA developers, and SOC firmware engineers. The second is a tutorial on the importance of effective marketing and should appeal to a broad range of DAC attendees that wish to better understand this aspect of business success. As in the past, the goal of the DAC tutorials is to provide practical, usable, and up-to-date knowledge that attendees can immediately apply in their jobs or studies.
MONDAY TUTORIALS
• ESL Design and Virtual Prototyping of MPSOCs
• Low-Power from A to Z
• Marketing of Technology - The Last Critical Step
FRIDAY TUTORIALS
• 3-D: New Dimensions in IC Design
• Advancing the State-of-the-Art in Analog Circuit Optimizers
• Best Practices for Writing Better Software
• SystemC for Holistic System Design with Digital Hardware, Analog Hardware, and Software
Detailed conference and exhibition information is now available online: www.dac.com.
Register today!
QUESTIONS? Call +1-303-530-4333
Sponsored by:
Sponsored by:
PANELSThis year’s DAC panels cover nearly every aspect of the design flow. The panel sessions start off with a look to the future by a wide range of leaders from the semiconductor industry. The other seven panels have something for everyone. Panels will explore the future of TSV/3D technology, the current state of high-level synthesis, different approaches to addressing process variability, the future of low-power design methodologies and how to bridge pre-silicon verification/post-silicon validation. One panel will also take a look at what is needed for an always-connected car. Finally, if you’ve wondered what cloud computing is all about, a panel will explore how cloud computing fits in with the EDA industry.
TUESDAY, JUNE 15
• EDA Challenges and Options: Investing For the Future
• Bridging Pre-Silicon Verification and Post-Silicon Validation
• Who Solves the Variability Problem?
WEDNESDAY, JUNE 16
• 3-D Stacked Die: Now or the Future?
• Does IC Design Have a Future in the Clouds?
• What’s Cool for the Future of Ultra Low-Power Designs?
THURSDAY, JUNE 17
• Designing the Always-Connected Car of the Future
• Joint User Track Panel - (Session 8UB) - What Will Make Your Next Design Experience a Much Better One?
• What Input Language is the Best Choice for High-Level Synthesis (HLS)?
SPECIAL SESSIONSSpecial sessions will deal with a wide variety of themes such as progress in networks-on-chip research, virtualization for mobile embedded devices, challenges in analog modeling, introduction to cyber-physical systems, design for reliability, designing resilient systems from unreliable components, a holistic view on energy management – cell phones to power grids and post-silicon validation. Leading research and industry experts will present their views on these topics.
TUESDAY, JUNE 15
• Post-Silicon Validation or Avoiding the $50 Million Paperweight
• Virtualization in the Embedded Systems: Where Do We Go?
• Joint DAC/IWBDA Special Session - Engineering Biology: Fundamentals and Applications
WEDNESDAY, JUNE 16
• A Decade of NOC Research - Where Do We Stand?
• The Analog Model Crisis - How Can We Solve It?
• Design Closure for Reliability
THURSDAY, JUNE 17
• WACI: Wild and Crazy Ideas
• Cyber-Physical Systems Demystified
• Computing Without Guarantees
• Smart Power: From your Cell Phone to your Home
WORKSHOPSSUNDAY, JUNE 13
• DAC Workshop on Synergies between Design Automation & Smart Grid
• Multiprocessor System-On-Chip (MPSOC): Programmability, Run-Time Support and Hardware Platforms for High Performance Applications at DAC
• DAC Workshop on Diagnostic Services in Network-On-Chips (DSNOC) - 4th Edition
MONDAY, JUNE 14
• IWBDA: International Workshop on Bio-Design Automation at DAC
• DAC Workshop on “Mobile and Cloud Computing”
• DAC Workshop: More Than Core Competence...What it Takes for Your Career to Survive, and Thrive! Hosted by Women in Electronic Design (WWED)
COLOCATED EVENTSFRIDAY, JUNE 11
• IEEE International High-Level Design Validation and Test Workshop (HLDVT 2010)
SUNDAY, JUNE 13
• International Symposium on Hardware-Oriented Security and Trust (HOST)
• 8th IEEE Symposium on Application Specific Processors (SASP 2010)
• Design for Manufacturability Coalition Workshop - “A New Era for DFM”
• IEEE/ACM 12th International Workshop on System-Level Interconnect Prediction (SLIP)
• North American SystemC Users Group (NASCUG 13 Meeting)
• System and SOC Debug Integration and Applications
MONDAY, JUNE 14
• 4th IEEE International Workshop on Design for Manufacturability & Yield (DFM&Y)
• Choosing Advanced Verification Methods: So Many Possibilities, So Little Time
• Advances in Process Design Kits Worshop
TUESDAY, JUNE 15
• ACM Research Competition
• NASA/ESA Conference on Adaptive Hardware and Systems (AHS-2010)
THURSDAY, JUNE 17
• IEEE/ACM International Symposium on Nanoscale Architectures (NANOARCH’10)
FRIDAY, JUNE 18
• 19th International Workshop on Logic & Synthesis (IWLS)
Accelicon Technologies, Inc. ACCIT - New Systems Research ACE Associated Compiler Experts bv Agilent Technologies Agnisys, Inc. Aldec, Inc. Altair Engineering Altos Design Automation Amiq Consulting S.R.L. AnaGlobe Technology, Inc.Analog Bits Inc. Apache Design Solutions, Inc. Applied Simulation Technology Artwork Conversion Software, Inc. ASIC Analytic, LLC ATEEDA Atoptech Atrenta Inc. austriamicrosystems AutoESL Design Technologies, Inc. Avant Technology Inc. Avery Design Systems, Inc. Axiom Design Automation BEEcube, Inc. Berkeley Design Automation, Inc. BigC Blue Pearl Software Bluespec, Inc. Breker Verification Systems Cadence Design Systems, Inc. Calypto Design Systems Cambridge Analog Technologies CAST, Inc. ChipEstimate.com Ciranova, Inc. CISC Semiconductor Design+Consulting GmbH ClioSoft, Inc. CMP CoFluent Design Concept Engineering GmbH Coupling Wave Solutions CST of America, Inc. DAC Pavilion Dassault Systemes Americas Corp. DATE 2011 Denali Software, Inc. Design and Reuse Dini Group DOCEA Power Dorado Design Automation, Inc. Duolog Technologies Ltd. E-System Design EDA Cafe-IB Systems EDXACT SA Entasys Inc. Enterpoint Ltd. EVE-USA, Inc. Exhibitor Forum ExpertIO, Inc. Extension Media LLC Extreme DA FishTail Design Automation, Inc. Forte Design Systems Gary Stringham & Associates, LLC GateRocket, Inc. GiDEL Global Foundries Gradient Design Automation
Helic, Inc. Hewlett-Packard Co. HiPEAC IBM Corp. IC Manage, Inc. ICDC Partner Pavilion & Stage IMEC - Europractice Imera Systems, Inc. Infotech Enterprises iNoCs Interra Systems, Inc. Jasper Design Automation, Inc. Jspeed Design Automation, Inc. JTAG Technologies Laflin Limited Legend Design Technology, Inc. Library Technologies, Inc. Lynguent, Inc. Magillem Design Services Magma Design Automation, Inc. Magwel NV MathWorks, Inc. (The) MentaMentor Graphics Corp. Mephisto Design Automation Methodics LLC Micro Magic, Inc. Micrologic Design Automation, Inc. Mirabilis Design Inc. Mixel, Inc. MOSIS MunEDA GmbH Nangate NextOp Software, Inc. Nusym Technology, Inc. Oasys Design Systems, Inc. OneSpin Solutions GmbH OptEM Engineering Inc. OVM World Physware, Inc. PLDA POLYTEDA Software Corp. Progate Group Corp. Prolific, Inc. Pulsic Inc. R3 Logic Inc. Rapid Bridge, LLCReal Intent, Inc. Reed Business Information RTC Group - EDA Tech Forum Runtime Design Automation Sagantec Sapient Systems Satin IP Technologies Seloco, Inc. Semifore, Inc. Si2 Sigrity, Inc. Silicon Design Solutions Silicon Frontline TechnologySKILLCAD Inc. Solido Design AutomationSonnet Software, Inc. Springer SpringSoft, Inc. StarNet Communications Synapse Design Synchronicity - see Dassault Systèmes
Synfora, Inc. Synopsys, Inc. Synopsys, Inc. - Standards Booth Synopsys-ARM-Common Platform Innovation SynTest Technologies, Inc. Tanner EDA Target Compiler Technologies NV Teklatech Tela Innovations Tiempo TOOL Corp. True Circuits, Inc. TSMC TSMC Open Innovation Forum, Apache TSMC Open Innovation Forum, Cadence TSMC Open Innovation Forum, eSilicon TSMC Open Innovation Forum, Helic, Inc.TSMC Open Innovation Forum, IntegrandTSMC Open Innovation Forum, LorentzTSMC Open Innovation Forum, Magma TSMC Open Innovation Forum, Mentor TSMC Open Innovation Forum, MoSys TSMC Open Innovation Forum, SolidoTSMC Open Innovation Forum, SpringSoft TSMC Open Innovation Forum, Synopsys TSMC Open Innovation Forum, Tela Innovations TSMC Open Innovation Forum, Virage Logic TSSI - Test Systems Strategies, Inc. Tuscany Design Automation, Inc. UMIC Research Centre Uniquify, Inc. Univa UD Vennsa Technologies, Inc. Verific Design Automation Veritools, Inc. WinterLogic Inc. X-FAB Semiconductor Foundries XJTAG XYALIS Z Circuit Automation Zocalo Tech, Inc.
Orange text denotes a new exhibitor
EXHIBITOR LIST (AS OF APRIL 12, 2010)
EXHIBITOR FORUM DAC is continuing the popular Exhibitor Forum again this year. The Exhibitor Forum provides a theater on the exhibit floor where exhibitors present focused, practical technical content to attendees. The presentations are selected by an all-user Exhibitor Forum Committee chaired by Magdy Abadir of Freescale Semiconductor, Inc. Each session is devoted entirely to a specific domain (e.g., verification or system-level design) and consists of presentations from three companies.
The Exhibitor Forum is in Hall B in Booth 1684. Topics include: System-Level Design/Embedded Software, Physical Design and Sign-Off, Verification, Power Management/Signal Integrity, Analog/Mixed-Signal and RF, Design for Manufacturability, Intellectual Property Cores, Design for Test and Manufacturing Test, Package Design, and Silicon Validation and Debug.
DAC PAVILIONThe popular DAC Pavilion is located in Hall C in Booth 694. The DAC Pavilion will feature 17 presentations on business and technical issues.
MONDAY, JUNE 14
• Gary Smith on EDA: Trends and What’s Hot at DAC
• The Multiplier Effect: Developing Multi-Core, Multi-OS Applications
• Career Outlook: Job Market 2010
• Outsourcing...!@#$*&!!?
• EDA Heritage - Meet Verilog Inventor Dr. Moorby and Formal Verification Pioneer Prof. Bryant
• A Conversation with the 2010 Marie Pistilli Award Winner
TUESDAY, JUNE 15
• Hogan’s Heroes: What Design and Lithography Nightmares will 22nm Bring?
• Everyone Loves a Teardown (ARM)
• Is the FPGA Tool Opportunity an Oasis or a Mirage?
• 28nm and Below: SOC Design Ecosystem at a Crossroad
• Hot and SPICEy: Users Review Different Flavors of SPICE and Fast SPICE
WEDNESDAY, JUNE 16
• Lucio’s Litmus Test: Is Your Start-Up Ready for the 21st Century?
• IP Commercialization: Beyond the Code
• Everyone Loves a Teardown (Virage Logic)
• High-School Panel: You Don’t Know Jack!
• Analog Interoperability: What’s the ROI?
• SOC Verification: Are We There Yet?
THE IC DESIGN CENTRAL PARTNER PAVILION—PUTTING MORE DESIGN INTO DACThe IC Design Central Partner Pavilion, located in Hall B, stage #1710, brings together vendors supplying products and services that address many of the critical design functions necessary to produce working silicon on time and on budget. Companies from all areas of the design and product development process—EDA, Foundry, IP, Design Services, Assembly/Package, Test, and System Interconnect—must cooperate to offer integrated front-to-back solutions that ensure first-time-successful silicon and predictable time-to-market. Visit the ICDC Partner Pavilion and find design flows and solutions needed to create today’s challenging designs.
The ICDC Partner Pavilion is a combination of exhibit booths and 30-minute presentations by each participating vendor. The combination of product displays in the exhibits and technical product presentations in the ICDC Theater offers attendees an in-depth look into flows and methodologies from vendors featuring a variety of products and services for the entire design ecosystem.
CURRENT PARTICIPATING ICDC EXHIBITORS INCLUDE:
EXHIBIT-ONLY PASSRegister for an exhibit-only pass and receive admission to all days of the exhibition, all Keynotes, all DAC Pavilion and Exhibitor Forum sessions, the IC Design Central Partner Pavilion, plus the Tuesday night DAC party, and a T-shirt —all for $50 when you register by May 17.
EXHIBITION HOURSMONDAY, JUNE 14 - WEDNESDAY, JUNE 169:00am - 6:00pm
Altair Engineering
Amiq Consulting S.R.L.
ASIC Analytic, LLC
Avant Technology Inc.
BEEcube, Inc.
Cambridge Analog Technologies
CoFluent Design
CISC Semiconductor Design
& Consulting
Enterpoint Ltd.
ExpertIO, Inc.
Gary Stringham & Associates, LLC
IBM Corp.
iNoCs
Progate Group Corp.
R3 Logic Inc.
TSSI - Test Systems Strategies, Inc.
X-FAB Semiconductor Foundries
Zocalo Tech, Inc.
EXHIBITIONThe 47th DAC exhibition is located in Halls B and C of the Anaheim Convention Center.
Visit the DAC exhibition for an in-depth view of new products and services from nearly 200 vendors spanning all aspects of the electronic design process, including EDA tools, IP cores, embedded system and system-level tools, as well as silicon foundry and design services.
Sponsored by:
REGISTRATION OPTIONS: Internet registration is open through June 18. Mail/fax registrations are accepted through June 8.
FULL CONFERENCE REGISTRATION includes: access to all three days of the Technical Sessions, User Track Sessions, Embedded/SOC Enablement Day, access to the Exhibition, Monday through Wednesday, the 47 Years of DAC DVD Proceedings and the Tuesday Night Party.
STUDENTS FULL CONFERENCE REGISTRATION IEEE MEMBER OR ACM MEMBERA special student rate applies to individuals who are members of ACM or IEEE and are currently enrolled in school. Students must provide a valid ACM or IEEE student membership number and a valid student ID. ACM/IEEE Student registration includes: all three days of the Technical Conference, Embedded/SOC Enablement Day, access to the Exhibition, Monday through Wednesday, the 47 Years of DAC DVD Proceedings and the Tuesday Night Party.
ONE/TWO DAY REGISTRATION INCLUDES: include the day(s) you select for the Technical Conference, access to the Exhibition, User Track (UT) Sessions, Monday through Wednesday, and the “47 Years of DAC” DVD Proceedings.
EXHIBIT-ONLY REGISTRATION allows admittance to the Exhibition, Monday through Wednesday and includes the Tuesday Night Party.
USER TRACK SESSIONS registration includes entrance to the Exhibition, Monday through Wednesday and all Keynotes. User Track Sessions are included in the Full Conference registration and the One-/Two-day registration on the day(s) attending the technical conference.
MANAGEMENT DAY registration for this event includes entrance to the Exhibition, Monday through Wednesday, and all Keynotes.
TUTORIALS are offered on Monday, June 14 and Friday, June 18. There is one quarter-day tutorial, two half-day tutorials, and four full-day tutorials. The full-day tutorial registration fee includes: continental breakfast, lunch, refreshments and tutorial notes. The half-day tutorial registration fee includes: continental breakfast, refreshments and tutorial notes. The quarter-day tutorial registration fee includes: refreshments and tutorial notes.
EMBEDDED/SOC ENABLEMENT DAY is a day-long track of sessions dedicated to bringing industry stakeholders together in one room to shed light on where SOC design is headed. The day is comprised of presentations from leading SOC enabling sectors, including embedded processors, embedded systems, EDA, FPGA, IP, foundry, and design services.
Advance Rate Received by May 17
Late/On-site Rate Received After May 17
Full Conference $475 $595 $230 $570 $695 $295
One-Day Only (Tue., Wed., Thurs.) $325 $325
Two-Day Only (Tue, Wed., Thurs.) $525 $525
Exhibit-only access all days (Mon. - Wed.)
$50 $95
Monday Exhibit-only (Monday) FREE FREE
Management Day (Tuesday) $95 $95
Embedded/SOC Enablement Day $95 $95
Full-day $300 $400 $200
Half-day $180 $240 $120
Quarter-day $100 $130 $80
User Track Sessions $185 $240
DAC Workshop on Diagnostic Services in Network-on-Chips (DSNOC) - 4th Edition
Sunday, June 13
$150 $195Multiprocessor System on Chip (MPSOC): Programmability, Run-Time Support and Hardware Platforms for High Performance Applications at DAC
DAC Workshop on Synergies between Design Automation & Smart Grid
DAC Workshop: More Than Core Competence...What it Takes for Your Career to Survive, and Thrive! Hosted by Women in Electronic Design (WWED)
Monday, June 14
FREE up to 100 attendees
DAC Workshop on “Mobile and Cloud Computing”
$150 $195
International Workshop on Bio-Design Automation at DAC (IWBDA)
Monday, June 14 & Tuesday, June 15
$230 $305
ACM/IEEE Member
Non-member
ACM/IEEE Student Member
Student Non-member
WORKSHOPS/COLOCATED EVENT registration also includes entrance to the Exhibition, Monday through Wednesday.
Visit the DAC website for online registration, complete conference and exhibition details, travel and hotel reservations and information on visiting Anaheim at www.dac.com.
CANCELLATION/REFUND POLICY:Written requests for cancellations must be received in the DAC office by Monday, May 17, 2010 and are subject to a $25.00 processing fee. Cancellations received after May 17, 2010 will NOT be honored and all registration fees will be forfeited. No faxed or mailed registrations will be accepted after June 8, 2010.
Telephone registrations are not accepted!
Faxed or mailed registrations without payment will be discarded.
Dates Advance Registration Late/Onsite Registration
IEEE International High Level Design Validation and Test Workshop 2010 (HLDVT) Fri. June 11 - Sat., June 12
$350 $450 $250 $250 $450 $575 $300 $300
IEEE/ACM 12th International Workshop on System Level Interconnect Prediction (SLIP) Sun., June 13
$250 $320 $200 $200 $300 $370 $250 $250
Design for Manufacturability Coalition Workshop FREE
IEEE International Symposium on Hardware-Oriented Security and Trust (HOST) Sun., June 13 - Mon., June 14
$300 $375 $150 $190 $360 $450 $180 $225
8th IEEE Symposium on Application Specific Processors (SASP 2010) $315 $410 $190 $190 $420 $530 $245 $245
4th IEEE International Workshop on Design for Manufacturability & Yield (DFM&Y)Mon., June 14
$150 $200 $100 $150 $200 $250 $130 $200
Advances in Process Design Kits Workshop FREE
NASA/ESA Conference on Adaptive Hardware and Systems (AHS-2010) Tue., June 15 - Fri., June 18
$560 $560 $410 $410 $690 $690 $510 $510
IEEE/ACM International Symposium on Nanoscale Architectures (NANOARCH’10)
Thurs., June 17 & Fri., June 18
$285 $375 $225 $300 $350 $440 $240 $315
Register Online by May 17 and Save!
Visit Our Website!
More of everything you’ve come to expect from Chip Design:
News
Technology Trends
Design Centers
Blogs
iDesign
Focus Report
Commentary
Technical Papers
Resource Catalogs and Guides
Email Newsletters
ChipDesignMag.com
Dedicated to the information needsof the IC design market
10 • April / May 2010 Chip Design • www.chipdesignmag.com
IN THE NEWSBy Jim Kobylecky, Managing Editor
SANDEEP VIJ NAMED CEO OF MIPS TECHNOLOGIESMIPS Technologies, Inc. has appointed
Sandeep Vij as president, chief executive
officer and director. Mr. Vij brings to
MIPS Technologies more than 20 years
of senior-level management and marketing
experience in the semiconductor industry. Prior to joining
MIPS Technologies, he was vice president and general
manager of the Broadband and Consumer Division of
Cavium Networks. Mr. Vij was had senior roles with Xilinx
Inc. and Altera. He is a graduate of General Electric’s Edison
Engineering Program and Advanced Courses in Engineering
and holds an MSEE from Stanford University and a BSEE
from San Jose State University.
THOMAS KAILATH WINS BBVA FOUNDATION AWARD FOR CHIP MINIATURIZATIONA BBVA Foundation Frontiers of Knowledge Award went to
engineer and mathematician Thomas Kailath, (Pune, India,
1935), the Hitachi America Professor of Engineering at
Stanford University, as author of a mathematical development
enabling the production of increasingly small size chips.
Kailath has invented methods to pattern integrated circuits
with components finer even than the lightwaves used in
their production. "I was able to see the opportunities and
enter new fields because I learned to use my students as
intelligence amplifiers," says Kailath. "So I regard this prize as
a tribute also to them, to their brilliance and dedication. No
comparable award scheme reserves a category for Information
Technologies. At 400,000 euros, it is the largest monetary
award in the ICT field.
XMOS APPOINTS CHARLES COTTON INTERIM CEOXMOS has appointed Charles Cotton
as a board member and interim Chief
Executive Officer. Cotton's deep experience
in leading and managing semiconductor
and software companies include his roles
as Executive Chairman of GlobespanVirata Inc. and CEO
of Virata Corp. before its merger with Globespan. He is
currently a Director of semiconductor companies Solarflare
and Staccato and two other companies in California. His UK
directorships include Cambridge Enterprise, the University of
Cambridge organization responsible for spin-outs, licensing
and consulting activities. Cotton was a Board member of
digital maps supplier Tele Atlas prior to its acquisition by
TomTom.
ESILICON NAMES AJAY LALWANI VP OF STRATEGIC SOURCINGeSilicon appointed of Ajay Lalwani as vice president, strategic
sourcing. In this role, Lalwani is responsible for expanding
eSilicon's capabilities in strategic sourcing across eSilicon's
entire global semiconductor supply chain. Additionally,
Lalwani will develop and manage strategic alliances with key
IP, EDA, photomask, wafer, assembly and test suppliers He
has over 22 years of experience in the industry and has gained
extensive insights into how business strategy, sales, marketing
and organizational development impact supply chains.
Lalwani has a BSEE and MBA from Santa Clara University.
CHIL SEMICONDUCTOR BOARD ADDS INTEL VP THOMAS MACDONALDThomas R. Macdonald has joined
the board of CHiL Semiconductor
Corporation. Mr. Macdonald has held
a range of general management and
high level strategic microprocessor and
platform marketing management positions since joining
Intel in 1988. Mr. Macdonald received his bachelor's degree
in mechanical engineering from Stanford University and his
MBA from the Kellogg Graduate School of Management,
Northwestern University.
DR. LEON O. CHUA RECEIVES ISQED QUALITY AWARD,The International Society for Quality Electronic Design
announced the winner of the prestigious 2010 ISQED
Quality Award (IQ-Award), Dr. Leon O. Chua of the
Electrical Engineering and Computer Sciences Department,
University of California, Berkeley. Dr. Chua is well known
as a pioneer in three major research areas, nonlinear circuits,
chaos and cellular neural networks. His work in these
areas has been recognized internationally by major awards,
including 12 honorary doctorates from major universities in
Europe and Japan and seven USA patents.
People in the News
Chip Design • www.chipdesignmag.com April / May 2010 • 11
IN THE NEWSBy Jim Kobylecky, Managing Editor
TRIDENT MICROSYSTEMS SELECTS DR. J. DUANE NORTHCUTT AS CTOTrident Microsystems, Inc. has named Dr. J. Duane
Northcutt as Chief Technology Officer (CTO). Prior to
joining Trident, Dr. Northcutt was with Silicon Image, Inc.
and was a Distinguished Engineer at Sun Microsystems.
Earlier, Mr. Northcutt was a member of the research faculty
at Carnegie Mellon University's School of Computer Science.
He currently holds over twenty patents. Dr. Northcutt
received both a Master's of Science degree and a Ph.D. degree
in computer and electrical engineering from Carnegie-Mellon
University.
ROB ROY JOINS ATRENTA EXECUTIVE TEAMAtrenta Inc. has selected Dr. Rob Roy to
be chief of business development. Dr. Roy
was a co-founder and chief technology
strategist of Mobilian Corporation. His
previous experience includes engineering
and management positions at Intel, NEC, GE, and AT&T
Bell Labs. Dr. Roy has published over 50 research papers in
prestigious journals and international conferences, including
three highly-prestigious Best Paper Awards. He holds 15
patents. He earned his M.S. and Ph.D. degrees in Electrical
& Computer Engineering from the University of Illinois at
Urbana-Champaign.
MAGMA NAMES ALOK MEHROTRA MANAGING DIRECTOR OF INDIAMagma Design Automation Inc. has
appointed Alok Mehrotra as managing
director of the company's India
operations. More than 30 percent of
Magma's worldwide workforce operates
in the company's Bangalore, Mumbai and Noida facilities.
Mehrotra had previously worked for the company as director
of Asia-Pacific Sales from 2001 to 2005, when he established
operations in India, Singapore, Malaysia and Australia.
Mehrotra holds an MBA from Santa Clara University;
an M.S. in electrical engineering from State University of
New York at Stony Brook; and a B.S. degree in electronics
& communication engineering from Manipal University in
Karnataka, India.
VERILAB PROMOTES JL GRAY TO VICE PRESIDENTVerilab Ltd. has promoted JL Gray to
the position of vice president, reporting
directly to chief executive officer, Tommy
Kelly. JL has contributed to the EDA
industry as Verilab’s representative on
the Accellera Verification IP Technical Subcommittee.
He is also the author of “Cool Verification,” a blog about
hardware verification from a consultant’s perspective. He has
worked extensively on the application of social media to the
EDA industry as a means of fostering collaboration in the
wider engineering community. JL has a BSEE from Purdue
University in West Lafayette, Indiana.
CYCLEO HIRES FRANÇOIS SFORZA AS VP OF SALESCycleo SAS has named François Sforza
to the position of Cycleo VP of Sales.
Sforza brings more than 20 years of Sales
& Marketing management experience in
wireless and semiconductors companies.
Prior to joining Cycleo, he was Regional Manager for Europe
at Wipro Technologies, providing Global Solution Services
to the Telecom, Semiconductor, Consumer and Automotive
Industries. François Hede, Cycleo’s CEO, said that “François
has a proven track record in the Wireless and Semiconductor
markets, as well as developing an ecosystem with partners &
end-users to help make our customers successful”.
MAGMA APPOINTS NORIAKI KIKUCHI PRESIDENT OF MAGMA KKMagma Design Automation Inc. has
named Noriaki Kikuchi president of
Magma KK, Magma’s Japanese subsidiary.
Kikuchi has more than 30 years experience
in electronic design automation and other
technology industries. With a history of sales and senior
management positions showing increasing responsibilities,
Kikuchi most recently was president of Japan operations for
Brion Technologies Inc., a subsidiary of ASML. Previously
he was president of Tera Systems Japan, and held senior sales
and field operations positions with Synopsys Japan and Seiko
Instruments Inc. Kikuchi holds a Bachelor of Arts degree in
management from Chuo University in Tokyo.
12 • April / May 2010 Chip Design • www.chipdesignmag.com
BEHIND THE NUMBERS
What is the functional make-up of today’s System-on-Chip
(SoC) designs? Unlike the past, when SoC were dominated
by digital logic and memory cores, today’s devices contain
a more balanced set of functional blocks. This viewpoint is
supported by the findings from a recent survey of the Chip
Design magazine readership (see Figure). The response was
surprising in that Digital Logic design edged out Analog
and Mixed Signal (AMS) by a slight margin – 6 percent.
Further, embedded processor cores edged out memory in
terms of the type of functional blocks utilized in SoC.
The shrinking cost, improved performance and lower power of
embedded processor, married with the strong growth in consumer
and mobile devices, is one reason for the increased presence of
embedded cores. A separate survey question concerning the type
of processor IP used in SoC designs revealed ARC as the leader,
followed by ARM, MIPS, Intel and others.
Interestingly, when survey respondents were asked about
third-party IP usage for their SoC designs, they selected
AMS blocks as being dominate, followed in a distant second
by embedded processor, memory and the digital logic IP.
These trends complement the move toward more embedded
processor designs and the increasing need for connectivity and
sensors in the growing market for consumer and mobile devices.
SoCs Move beyond Digital and Memory Blocks
By John Blyler, Editor-in-Chief
IC Functional Block
Analog andMixed
Signal, 59,20%
DigitalLogic, 77,
26%
EmbeddedProcessors,
49, 17%
Input-Output, 40,
14%
Memory,47, 16%
RF/Wireless, 22, 7%
Blank = 18 or 6%Total Responses = 312 (Multiple Choice)Total Respondents = 116
Dual Core Embedded Processors Bring Benefits And ChallengesBy John BlylerThe embedded processor market has now fully embraced the
multicore world with the recent introduction of the dual core
option for Intel’s Atom devices. Dual-core embedded proces-
sors offer designers many new benefits while presenting new
challenges. How will the multicore option affect low power
designs, virtualization, and single-threaded legacy software?
Will these devices lead to more connectivity? Is the embedded
processor market looking like the ASSP market of the future?
To answer these questions, Low-Power Engineering talked
with Jonathan Luse, Director of Marketing for the Low-Pow-
er Embedded Products Division of Intel.
LPE: How does dual-core affect power consumption?
Luse: It’s best to think of the Atom as roughly split into two
vectors–performance and power. The performance vector is
a little less power constrained and a little more performance
oriented, but still low power compared to Intel’s Core family
of processor. The other major vector is low power. At the
winter Embedded World Conference in Nürnberg, Germany,
we introduced our entry performance level processors, which
included the dual-core option at about 13watts thermal
design power (TDP) to 5.5 watts for the single-core kit at
1.6kHz. This was designed to have a little more tolerance for
power, with the expectation that Input/Output (IO) interface
and performance would be increased over time.
To read more, please visit: lpdcommunity.com
,
Is your company a missing piece in the EDA & IP industry?The EDA Consortium Market Statistics Service (MSS) is the timeliest and most detailed market data available on EDA & IP. Top companies have been participating in the MSS since 1996, but no report is complete without ALL the pieces.
If your company is not contributing to the MSS, it is a missing piece in the puzzle. Contribute your company’s data to the MSS and receive a complimentary Annual MSS Executive Summary.
What’s in it for your company?• Worldwide statistics on EDA & IP revenue• Ability to size your market• Six years of market data for trends analysis• The opportunity to make a positive difference in the industry
For details contact:[email protected](408) 287-3322www.edac.org
Receive a Complimentary EDA & IP Industry Market Report
14 • April / May 2010 Chip Design • www.chipdesignmag.com
FOCUS REPORTBy Ed Sperling, Contributing Editor
The consolidation of intellectual property from small
developers to large players with integrated IP blocks is
accelerating. Large IP companies are now developing
integrated suites that are pre-tested for specific vertical
markets, and new companies are sprouting up to make it
easier to put even broader collections of IP together in
meaningful ways.
It’s difficult to tell whether the trend is being driven more
by the IP vendors or pulled through by chip developers
looking to cut costs—or whether it builds upon the stamp
of approval by foundries for certain pieces of IP. The net
effect, however, is the creation of subsystems and partial
platforms that are one step below reference platforms.
“A reference design suggests a complete solution,” said Eric
Schorn, vice president of marketing for ARM’s processor
division. “Customers don’t want us to go that far. But we
are moving in a segment-oriented fashion. That’s the reason
we bought a graphics processor company. We are making a
processor along with a graphics socket for mobile phones
and set-top boxes.”
The company isn’t alone in recognizing the opportunity for
putting together more pieces of IP in very specific ways.
Virage Logic’s recent acquisitions of ARC and NXP’s IP
unit have positioned it to lead with integrated subsystems
in markets such as high-performance audio and video.
“You have to have a reference platform these days,” said
Yankin Tenurhan, vice president and general manager of
Virage’s ARC business unit. “ That’s not much different
from the good old days of silicon, though, when you needed
a complete solution and a full blown prototype. Philips,
NXP, Texas Instruments and ST all have demonstrator
chips for whatever you want on a cell phone. The same is
happening in the IP world.”
PUTTING TOGETHER THE PIECES
It’s not just the IP vendors that are putting together suites
of IP. Two startups are focused on making IP easier to
understand and integrate. Parallel Engines, which emerged
from stealth mode this week, is focused on organizing IP by
data mining pertinent information about everything from
power requirements to the interfaces and interconnects.
“ There are 12,000 pieces of IP out there, including 8,000
pieces of hard IP that are made by about 50 companies and
about 4,000 pieces of soft IP,” said George Janac, CEO of
Parallel Engines. “ The hard IP is already in FPGAs from
companies like Actel, Xilinx and Altera. You just need the
soft IP to make it work.”
Somewhat conveniently, Janac’s brother, Charlie, is the
CEO of Arteris, which makes network on chip technology
that can be used to glue together these IP blocks.
“A company may have one or two pieces of IP that are the
secret sauce and some software,” Charlie Janac said. “Why
not drop those into an FPGA and connect up the other
pieces of IP? Those two worlds are merging. We’re going to
see much more custom logic on an FPGA.”
Another company involved in bringing IP together is Silicon
IP, run by Kurt Wolf (formerly of TSMC), who said there’s
a disconnect between chipmakers and IP vendors that still
needs to be closed. “ The chip guys distrust the IP industry,”
Wolf said. “ There’s more integration of IP, but there’s still
a lack of confidence about how to choose, buy and license
IP.”
Wolf ’s company is focused more on bringing the two sides
together with better information and connecting the pieces
in an organized way.
THE FUTURE
All of these efforts—by both large IP vendors and
startups—are signs of just how important commercial
IP has become in chip development. What began with
embedded processors and standard memory designs has
evolved into a huge market that actually gained momentum
in the recent downturn.
Integrated IP Goes Vertical
Chip Design • www.chipdesignmag.com April / May 2010 • 15
FOCUS REPORTBy Ed Sperling, Contributing Editor
Outsourcing is gaining ground at every level of business,
even outside of the semiconductor world, but in the past
most of the gains have been in areas where there was little
value add. Outsourcing traditionally has been relegated to
commodity services. What’s changing is that IP now includes
areas that companies cannot do themselves in addition to
those they don’t want to do, as well as the extremely tedious
and time-consuming integration work that is necessary to
create a final product.
When most analysts predicted a massive growth in IP at
the beginning of the decade they were largely talking about
small, relatively unsophisticated IP blocks pieces that can
be put together by highly sophisticated companies. In the
future, the differentiation may be less around the technology
and more on getting very complex chips assembled and to
market faster for specific market segments.
Ed Sperling is Contributing Editor for
Embedded Intel® Solutions and the Edi-
tor-in-Chief of the “System Level Design”
portal. Ed has received numerous awards
for technical journalism.
WWW.si2.orgFeatured DownloadsFeatured DownloadsAt: www.si2.org/?page=1038
• 90+ Si2 Corporate Members
• All coalitions led by companieswho design ICs
• Board of Directors: AMD, ARM,Cadence, IBM, Intel, LSI, NationalSemiconductor, NXP, SequenceDesign, Synposys
Join the leaders in thesemi and EDA industry!
Far more than just Standards....Far more than just Standards....
......It’s about Adoption and ROI......It’s about Adoption and ROI
Slow Adoption for ESLBy Brian FullerIt’s been more than a decade since electronic system
level (ESL) abstraction started to gain traction in
EDA. It’s been more than a few years since the in-
dustry began to plan for the day when the benefits of
embracing C-language approaches to design descrip-
tion and validation would find designers churning out
massively complex and profitable designs while sitting
in lawn chairs sipping drinks with little pink umbrel-
las in them.
What happened?
Well, it’s still with us, but no one’s broken out the lawn
chairs just yet
To read more, please visit the System-Level Design com-
munity at sldcommunity.com
16 • April / May 2010 Chip Design • www.chipdesignmag.com
HAR
DW
AR
E-S
OFTW
AR
ED
ESIG
NPR
OTO
TYPIN
G
By Frank Schirrmeister, Synopsys
Prototyping Options for
Hardware/Software DevelopmentHow to choose the right prototype for pre-silicon software development
With software development playing an increasing role in
determining overall project effort and time-to-market for
consumer, wireless and automotive projects, pre-silicon software
prototyping has become a necessity to ensure that hardware and
software interact correctly with each other. However, choosing the
appropriate prototype for software development and hardware/
software co-design is not a trivial undertaking.
To understand the tradeoffs involved in this process, let’s examine the
concrete project development example shown in Figure 1. The upper
portion of the Gantt chart displays the timeline for the different project
phases, while the bottom portion shows the percentage of overall
project effort expended on all of these phases. Qualification of the IP
actually takes significant effort, as does the actual design management.
Relevant for the modeling issues is that verifying the RTL dominates
both the timeline and the actual effort. For this project, it took a total
of four quarters to get to verified RTL, and silicon prototypes were
not available until eight quarters into the project. The overall software
development took five quarters, so if it indeed had to wait for silicon
to be developed, the project would be significantly delayed.
Prototyping for software development can be done at different
stages, with various pros and cons. Although not reflected in Figure
1, previous-generation chips are often used for actual application
software development while the project is under way. Depending
on the number of changes from one chip generation to the next,
software can be developed on the previous device, and as soon as
new drivers are available, they can be replaced. If product capabilities
change significantly between product generations, this option
becomes less attractive, as the important new capabilities need to be
developed after updated drivers and silicon are available.
DIFFERENT PROTOTYPING OPTIONS AND THEIR AVAILABILITY
Available earliest in a project are virtual prototypes. They represent
fully functional software models of systems on chip (SoCs), boards,
I/Os and user interfaces, they execute unmodified production code,
and they run close to real-time with virtualized external interfaces.
They also offer the highest system visibility and control, including
multi-core debug. While virtual prototypes offer very high speed
(multiple tens of MIPS) when using so-called loosely timed models,
their limited timing accuracy often causes design teams used to
hardware prototypes to be skeptical of their value. The speed of
virtual platforms will degrade to the single-digit MIPS range or even
lower if users choose to mix in more timing-accurate software models.
However, because virtual prototypes are available earliest in the flow
if models are available, they are great for debugging and control, as
simulation provides almost unlimited insight, and they are also easy
to replicate. They offer true hardware/software co-development
capabilities because changes in the hardware can still be made if the
virtual prototype is available early enough in the design flow.
Variations of virtual platforms are so-called software development
kits (SDKs) – for example the iPhone SDK, which was downloaded
more than 100,000 times in the first couple of days of its availability.
While they offer most of the advantages of the standard virtual
prototypes, their accuracy is often more limited because they may
not represent the actual registers as accurately as virtual prototypes
but instead allow programming toward higher-level application
programming interfaces (APIs) and often require re-compilation
of the code to the actual target processor after users have verified
functionality on the host machine on which the SDK executes.
Available later in the design flow, but still well before silicon, FPGA
prototypes can serve as a vehicle for software development, as well.
They are fully functional hardware representations of SoCs, boards
and I/Os. They implement unmodified ASIC RTL code and run
at almost real-time speed, with all external interfaces and stimulus
connected. They offer higher system visibility and control than the
Figure 1: Concrete hardware software project development example.
(Source: Joint analysis by Synopsys and International Business Strategies)
Chip Design • www.chipdesignmag.com April / May 2010 • 17
ESL
HAR
DW
AR
E-S
OFTW
AR
ED
ESIG
NPR
OTO
TYPIN
G
Figure 2: Eight model characteristics for choosing prototyping solutions
actual silicon will provide later, but do not quite match the debug and
control capabilities of virtual platforms. Their key advantage is their
ability to run at high speed – multiple MIPS or even 10s of MIPS
– while maintaining RTL accuracy, but depending on the complexity
of the project, they will typically be available much later in the design
flow than virtual prototypes. Due to the complexity and effort of
mapping the RTL to FPGA prototypes, it is not really feasible to use
them before RTL verification has stabilized. Finally, once stable and
available, the cost of replication and delivery for FPGA prototypes is
higher than for software-based virtual platforms.
Emulation provides another hardware-assisted alternative to
enable software development. It differs from FPGA prototypes
in that it enables better automated mapping of RTL into the
hardware together with faster compile times, but the execution
speed will be lower and typically drop to the single-MIPS range
or below. The cost of emulation is also often seen as a deterrent to
replicating it easily for software development. Both emulation and
FPGA prototypes are limited when it comes to true hardware/
software co-development because at this point in the design flow,
the hardware is pretty much fixed, as RTL is almost verified.
Design teams will be very hesitant to change the hardware
architecture unless a major architecture bug has been found.
Finally, after the actual silicon is available, early prototype boards using
first silicon samples can enable software development on the actual
silicon. Once the chip is in production, very low-cost development
boards can be made available. At this point, the prototype will run at
real-time speed and full accuracy. Software debug is typically achieved
with specific hardware connectors using the JTAG interface and
connections to standard software debuggers. While prototype boards
using the actual silicon are probably the lowest-cost option, they are
available very late in the design flow and allow almost no head start
on software development. In addition, the control and debug insight
into hardware prototypes is very limited unless specific on chip
instrumentation (OCI) capabilities are made available. In comparison
to virtual prototypes, they are also more difficult to replicate – it is
much easier to provide a virtual platform for download via the
Internet then to ship a board and deal with customs, bring-up and
potential damages to the physical hardware.
WHICH PROTOTYPE SHOULD I CHOOSE?
So, how do users choose the appropriate prototype for early software
development? Several characteristics determine the applicability of
the chosen prototyping approach and the models it is built from.
Summarized in Figure 2, they fall into the following eight categories:
• Time of Availability: The later models become available in the
design flow compared to real silicon, the less their perceived
value to hardware/software developers will be.
• Execution Speed: Developers normally ask for the fastest
models available. Execution speed almost always is achieved by
omitting model detail, so it often has to be traded off against
accuracy.
• Accuracy: Developers normally ask for the most accurate models
available. The type of software being developed determines how
accurate the development method must be to represent the
actual target hardware, ensuring that issues are identified at
the hardware/software boundary. However, increased accuracy
requires simulating more detail, which typically means lower
execution speed.
• Production Cost: The production cost determines how easily a
model can be replicated for furnishing to software developers. In
general, software models are very cost-effective to produce and
can be distributed as soon as they are developed. Hardware-
based representations, like FPGA prototypes, require hardware
availability for each developer, often preventing proliferation to a
large number of software developers.
• Bring-up Cost: Any required activity needed to enable a models
outside of what is absolute necessary to get to silicon can be
considered overhead. The bring-up cost for virtual prototypes
and FPGA prototypes is often seen as a barrier to their use.
• Debug Insight: The ability to analyze the inside of a design,
i.e., being able to access signals, registers and the state of the
hardware/software design, is considered crucial. Software
simulations expose all available internals and provide the best
debug insight.
• Execution Control: During debug, it is important to stop the
representation of the target hardware using assertions in the
hardware or breakpoints in the software. In the actual target
hardware, this is very difficult – sometimes impossible – to
achieve. Software simulations allow the most flexible execution
control.
18 • April / May 2010 Chip Design • www.chipdesignmag.com
HAR
DW
AR
E-S
OFTW
AR
ED
ESIG
NPR
OTO
TYPIN
G
Figure 3: Combination of different prototyping options
• System Interfaces: It is often important to be able to connect
the design under development to real-world interfaces. While
FPGA prototypes often execute fast enough to connect directly,
development using virtualized interfaces of new standards, e.g.,
USB 3.0 can be done even before hardware is available.
The choice of prototype is often simplified to the trade-off between
speed and accuracy. In this context, the type of software to be
developed directly determines the requirements regarding how
accurately hardware needs to be executed.
• Application software can often be developed without taking the
actual target hardware accuracy into account. This is the main
premise of SDKs, which allow programming against high-level
APIs representing the hardware.
• For middleware and drivers, some representation of timing may
be required. For basic cases of performance analysis, timing
annotation to caches and memory management units may be
sufficient, as they are often more important than static timing of
instructions when it comes to performance.
• For real-time software, high-level cycle timing of instructions can
be important in combination with micro-architectural effects.
• For time-critical software – for example, the exact response
behavior of interrupt service routines (ISRs) – fully cycle-
accurate representations are preferred.
Given the above considerations, it comes as no surprise that none of
the prototyping techniques fits all applications. For users who need
to balance time of availability, speed and accuracy of prototypes,
combining different prototyping techniques offers a viable solution.
Figure 3 compares six different prototyping techniques and some of
their combinations using the example of an ARM-based platform
executing Linux connected to a USB 2.0 interface. Depending on the
use case, different combinations of TLM and signal-level execution
may be preferable. For example:
• For verification, the combination of transaction-level models
(TLM) with signal-level RTL offers quite an attractive speed-
up, and users have started to adopt this combination of mixed-
level simulation for increased verification efficiency. This use
model is effective even when RTL is not fully verified yet and
FPAG prototypes are not yet feasible.
• For software development, system prototypes, i.e. the
combination of virtual prototype using TLMs and FPGA
prototypes at the signal-level using standard interfaces like SCE-
MI, have become an attractive alternative for providing balanced
time of availability, speed, accuracy and debug insight for both
the hardware and software. This use case is most feasible once
RTL is mostly verified and the investment of mapping it into
FPGA prototypes is worth the return in higher speed.
Today, many companies already view prototyping as mandatory
to ensuring functional correctness of their designs and enabling
early software development. As this article has illustrated,
however, there is no “one size fits all” prototyping solution –
developers must select the approach that best meets their specific
project requirements. One thing is certain: with the trend toward
software continuing to escalate, implementing prototyping and
combinations of different prototyping techniques will gain even
greater importance for future design projects.
As director of product marketing at Synopsys, Inc.,
Frank Schirrmeister is responsible for the System-
Level Solutions products Innovator, DesignWare®
System-Level Library and System Studio, with
a focus on virtual platforms for early software
development. Prior to joining Synopsys, Frank
held senior management positions at Imperas, ChipVision, Cadence,
AXYS Design Automation and SICAN Microelectronics.
... there is no “one size fits all” pro-totyping solution – developers must select the approach that best meets their specific project requirements.
Chip Design • www.chipdesignmag.com April / May 2010 • 19
TR
AN
SACTIO
N-L
EVEL
MO
DELIN
GESL
By Lauro Rizzatti, EVE-USA
Making Abstraction PracticalA Vision for a TLM-to-RTL Flow
The ballooning size and complexity of system-on-chip
(SoC) designs has become an urgent driver of higher
levels of design abstraction. Just as it’s been a long time
since you could design electronic circuits one transistor at a
time, it’s now impossible to create an SoC one gate at a time.
Not only would your design time push you way beyond the
useful market window, but coordinating the roles of the
many members of your design team, from architect to tester,
would be impossible.
Transaction-level modeling (TLM) has provided a means
for starting designs at a more abstract level. But the path
from TLM down to physical implementation is far from
smooth. There are too many holes and inconsistencies in
the flow; each company has to invent something themselves
to get a useful result. The widespread use of third-party
intellectual property (IP) and the need to incorporate
software add entirely new dimensions to the problem. What
you need is a more unified approach to turning high-level
abstract design concepts into real chips.
MOVING UP A LEVEL
It’s hard to use the word “unified” when describing SoC
flows. Depending on the process node and performance
requirements, you have innumerable options involving
speed, power, and manufacturing yield optimization.
However, for a design described in RTL, it’s still relatively
straightforward to push the design through a flow and have
polygons come out the other end. That flow may make use of
involved scripts put together by clever CAD managers, but
variations within the digital domain are generally related to
optimization rather than actual behavior.
So even with these variations, the RTL-to-silicon flow is far
more predictable than what’s required to transform a more
abstract description into RTL. You simply can’t assemble an
SoC using a single behavioral description in one language.
• Architects need to be able to experiment with broad
ranges of functionality without having to specify gate-
level behavior. They should be able to make first-level
performance, power, and area tradeoffs.
• Verification must be achievable at an early stage without
gate- and cycle-level simulation.
• Architects and designers don’t specify in detail the
functionality of the entire SoC. Some blocks will be
designated for detailed custom design, but many will be
imported as IP, with their internal workings opaque.
• Software, executing in one or more processors on the
SoC, provides ever greater amounts of functionality.
Writing that software can take as long, or longer, than
designing the silicon platform on which it will run.
• Early validation of architecture and software often
requires emulation hardware, much of which uses
technology like FPGAs, which is far different from
what the end silicon will look like. You must therefore
be able to express the design at a level where it you can
target it at the emulator and at silicon without requiring
significant rework.
The result, even before taking into account any analog
circuitry you need to have on-chip, is a heterogeneous
amalgam of bits and pieces that you have to bring together
into a design. In the early stages of the planning, you may
designate some of the blocks as IP, with their functionality
either partially or completely known; you’ll mark others
for custom creation, and you won’t know their specific
behavior until someone actually does the design. You will
have some functionality written in RTL (even if bare-
bones), some in SystemC, and some in C/C++ or some
other software language.
This means that architects need to be able to pull the pieces
together at a “rough” level, making sure that everything plays
nicely together, or specifying the rules so that everything
will play nicely, and then dispatching the pieces for
implementation and integration. TLM provides a way for
architects to manage such high-level planning, but if the
designers that will implement custom blocks essentially
end up throwing away what the architects did and starting
their designs based on a paper spec, not only is work being
redone, but errors can be newly introduced. A flow that
connects the TLM work to the RTL work will reduce both
design time and the number of validation iterations.
20 • April / May 2010 Chip Design • www.chipdesignmag.com
ESL
TR
AN
SACTIO
N-L
EVEL
M
OD
ELIN
G
DIFFERENT LEVELS OF TLM
It’s inaccurate simply to talk about TLM as if it were a single
level of abstraction above RTL. Abstraction comes at the
cost of accuracy, and, depending on the task, you may need
to select different levels of abstraction in order to achieve
sufficient accuracy depending on what you’re trying to do.
The key to achieving this is the fact that TLM really deals
with interfaces: different blocks are plugged together, with
their behaviors abstracted and their interfaces interacting.
Accuracy boils down to the fidelity with which the interface
will model the finished block’s behavior. Greater accuracy
comes at the cost of longer verification times, and different
development phases will require different tradeoffs between
accuracy and verification time, as illustrated in Figure 1.
• Software engineers need the least fidelity; all they need
is for the block to function correctly. Timing is, more
or less, not an issue. This allows for highly abstracted
functional models, also known as virtual prototypes,
that can execute in the range of 10 – 100 million
instructions per second.
• Architects need a higher level of accuracy so that they
can confirm, for example, that bus-level transactions
occur properly. Here the level of “handshake” may be
sufficient; the actual number of clock cycles occurring isn’t
important. Such cycle-approximate models can execute
in the range of a million instructions per second.
• For more detailed verification of RTL blocks, designers
need to verify the cycle-accurate behavior of the
interfaces. This further slows the models down to around
100,000 cycles per second (which is more than an order
of magnitude slower than the “handshake” level since
we’ve gone from instructions-per-second to cycles-per-
second, and an instruction takes more than one cycle).
This means that simply having a single TLM model for a
block or piece of IP isn’t sufficient. Different blocks may
have different accuracy levels; you can’t just plug them
together and expect them to work. In fact, designers of
IP and custom blocks may need to develop multiple
interfaces to address the needs of different steps in the
design process. Exactly what those expectations should be
has not been standardized.
The cost of developing these different levels of TLM model
varies widely. Virtual prototypes are easiest because they’re
written in software, and therefore can take advantage of all
the tools available in the software world for verification and
debug. They also require that only the salient functionality
of the model be implemented. This means that, typically,
you can get virtual prototypes of common IP functions
from companies that don’t sell the actual implementation
IP. These companies focus their value on structuring the
models so that they’ll execute quickly and efficiently.
Cycle-approximate and cycle-accurate models require
much more work to build, and are much more closely tied
to the specific IP they model. Therefore, when purchasing
IP from a given vendor, you will typically get these models
from them, since only they know, and can model, the inner
workings of their secret circuits. Developing these models
can represent as much as 30% of the effort required to
create the RTL code itself.
DEFINING A FLOW
Having identified not one, but at least three different
kinds of TLM model that, in one form or another, define
functionality that will end up in silicon (or software on
silicon), the next obvious question is how to craft a flow
that, in the ideal, allows you to synthesize from the abstract.
It’s likely that such synthesis would be inefficient in its
early days, but that was also the case with RTL when logic
synthesis was new; eventually the tools have improve to
the point where only in rare cases would you countenance
doing a digital design at a level below RTL.
Before we can have a discussion of TLM synthesis, however,
we have to decide which of the tasks currently done at the
RTL level can be pushed up to a more abstract level. If we’re
eventually going to bury RTL in a flow in the way that EDIF
is buried today, then key manual steps currently done with
RTL have to be available at the TLM level along with the
new capabilities that TLM enables.Figure 1. Different model accuracy requirements
Chip Design • www.chipdesignmag.com April / May 2010 • 21
TR
AN
SACTIO
N-L
EVEL
MO
DELIN
GESL
The most important RTL task to push up is that of
optimization – especially for performance and power. It’s
well known that you get the biggest speed and power gains
if you optimize at the architecture level, where you can have
an impact of anywhere from fifty to hundreds of percent.
Circuit-level optimizations are good, but typically give you
a few tens of percent at best, and more and more of the
circuit-level tricks can be automated.
This means you need optimization tools that work at the
TLM level, along with models that provide estimates of
power and performance accurately enough to make the right
architectural tradeoffs.
IP evaluation is another task that you will have to manage
at the TLM level. While power and performance are a part
of that evaluation, you must also be able to confirm that
the IP you’re considering will play nicely with the rest of
the system, with as little wrapping as possible. You also
need to know that you can implement all the features you
need, and that you have to implement very few, if any, of
the features you don’t need.
Software evaluation is a newer task for the architect. Part
of the job may be actually checking out a specific piece of
software, but, to a large extent, the big problem is ensuring
that typical software can execute efficiently – you’re not
testing the software, you’re testing the system as it runs
software.
Having accomplished these tasks at the abstract level,
implementation can begin. There are really three different
elements to implementation:
• Creation of new blocks
• Assembling the blocks (newly created and IP)
• Creation of software
Functional verification then means
• Testing the architecture using cycle-approximate
models
• Checking out the new blocks using cycle-accurate
models and simulation
• Testing the assembly of the entire system. This
would actually happen in stages, starting with cycle-
approximate models and transitioning to cycle-
accurate where needed. Since full simulation of the
entire system is likely to take too long, hardware
emulation becomes a final way to test the integrated
logic of the entire design. Actual circuit simulation
might be used for very specific corner cases. Emulation
and simulation can also be run together to balance
speed and accuracy, either for hard-to-reach corner
cases or for early testing, where some models aren’t
available in abstract versions and so must be emulated
to keep performance up.
• Testing software using virtual prototypes, which tests
the software algorithms.
• Testing software on a hardware emulator, which tests
the system’s ability to execute the software.
This only makes sense if the work done early at the abstract
level can be used later to confirm the implementation
work. Today that would mean using the abstract models
to confirm the behavior of hand-generated blocks. If
TLM-level synthesis were available, then tools could be
used to confirm the correctness of that synthesis, much
the way equivalence checking was used to validate early
logic synthesis tools.
In order for this to work, however, if you have robust debug
capabilities that span the range of abstraction. If a signal in
an RTL block is misbehaving, that problem must ultimately
be correlatable to some higher-level model behavior. This
isn’t so hard when going from the specific to the abstract,
but going the other direction is harder: trying to correlate
a high-level model failure with a specific implementation
issue is tough because the high-level model has such specifics
abstracted out – by intent.
Debugging must also span different verification
methodologies. Where you are doing hardware emulation
and simulation together, for example, a debug methodology
has to recognize events and design elements on both the
simulated and emulated sides.
All of these requirements are hard enough to achieve
today in the purely digital domain. But SoCs increasingly
include significant analog functionality as well, and the
days of analog and digital ignoring each other are fast
disappearing. All of the flow elements involving modeling,
architecture, implementation, validation, and debug apply
equally – and raise even greater challenges – for analog.
Figure 2 shows what a complete TLM-to-RTL flow might
look like.
22 • April / May 2010 Chip Design • www.chipdesignmag.com
ESL
TR
AN
SACTIO
N-L
EVEL
M
OD
ELIN
G
The phases and steps can be described as follows:
• Architecture phase
• Create a system model, including a stimulus
environment, drawing from an IP library as much as
possible.
• Create new TLM models where needed.
• Simulate to validate the architecture and functionality
(including software) at the TLM level.
• Create a virtual prototype to give to the software
development team.
• Hardware design phase
• Automatically map IP blocks to RTL and generate the RTL
interconnect.
• Create RTL blocks either by synthesizing from TLM
models or by hand, with equivalence checking to ensure
that the generated RTL matches the TLM models.
• Perform full-chip RTL simulation for select corner
cases.
• Software design phase
• Create software, validating with the virtual prototype.
• Integration phase
• Perform complete hardware/software integration
validation using simulation and emulation.
• Connect to the physical design flow for implementation.
REQUIREMENTS FOR UNIFYING FLOWS
If such a flow is going to be possible without each company
defining its own proprietary version, some common elements
need to come together in the form of standards and ecosystem
offerings.
• IP modeling needs to be made consistent, with new
characteristics that will make evaluation easier. These
include interface standards, performance estimates of
key transactions, power estimation, and area estimation.
• Formalization of different TLM levels will help ensure
that users of IP can obtain appropriate models, and that
they will know what to expect when getting them.
• Tools need to be able to manage verification across
different levels of abstraction, allowing the assembly of
abstracted or partially-complete blocks with RTL-level
blocks, providing optimal speed and accuracy points.
• It’s currently possible to mix simulation and emulation, but
continued work is required to improve that integration as
well as to enable virtual prototype execution as part of a
coordinated environment.
• A unified debug environment is needed to ensure that
problems can be easily identified regardless of the level
of abstraction or the mode of verification, whether static
or dynamic.
These elements cross the property lines of a few different standards
and standards bodies. OSCI has owned the TLM specifications;
the SPIRIT Consortium, now a part of Accellera, has focused
on IP metadata; emulation interaction has been the domain
of the SCE-MI standard, also owned by Accellera. There is no
organization specifically focusing on the needs of debug.
Coordination and cooperation between the different standards
groups and sub-groups as well as between the companies
participating in the standards will be needed to provide all the
links to make this work. While some companies resist standards
out of fear of losing a competitive advantage, there are plenty
of opportunities to compete even in the face of a unified flow.
Each step of the flow will be challenged to provide the highest
performance, the greatest productivity, and the appropriate cost.
The industry as a whole will benefit by focusing innovation on
those areas, and, as the industry moves forward, participating
companies will have greater opportunities to reap the rewards.
At the same time, users must embrace the technology and
validate flows. It’s insufficient for tools providers simply to
support designs input in higher-level languages and synthesize
RTL. All of the pieces described above must be woven together
into methodologies that gain real traction with real users. Only
then can the TLM-to-RTL flow be considered reality.
Lauro Rizzatti is general manager of
EVE-USA. He has more than 30 years
of experience in EDA and ATE, where he
held responsibilities in top management,
product marketing, technical marketing
and engineering.
Figure 2. TLM-to-RTL Flow
Chip Design • www.chipdesignmag.com April / May 2010 • 23
FIR
MW
AR
ECO
-VER
IFIC
ATIO
N
SIM
ULATIO
N
By Marc Bryan and Barry Pangrle, Mentor Graphics
Avoid That Embarrassing Call to the Firmware
Vendor (and Other Tricks of Low-Power Verification)Simulation-based hardware/software co-verification lets you address power problems early and well.
Just how important is hardware/software co-verification
in low-power ASIC and SoC design and engineering?
You might ask the large semiconductor company* that a few
years back designed a device per the specs of a significant
customer, which assembled and sold smartphones. The
specs – that a varied combination of functions could
execute concurrently without exceeding a certain power
budget, measured in milliwatts – were fairly typical for the
low-power realm.
When the customer received the silicon, however, its
engineers struggled to get the device to behave as advertised.
Despite their best efforts to string together functions that
should have been well within the device’s limits, their
applications kept exceeding the power budget.
It should be said that the semiconductor company deserves
heapings of kudos for sticking around to help. After realizing
it didn’t have tools or expertise to do extensive hardware/
software co-verification, the company wound up hiring an
entire firmware team. After much effort, expense, and, most
significantly, delay to the customer’s product, these contract
coders put together software infrastructure that bridged
the ASIC’s underlying power-management features to the
customer’s skills and objectives for the new product.
The story is not unusual. Given the increasing complexity of
ASICs and SoCs, it’s no longer enough for semiconductor
companies to focus on silicon and deliver meager amounts
of diagnostic software as an afterthought. This is especially
true where power-management looms large, as it does in
just about any device with batteries, a market segment that
seems poised for a major rebound. The worldwide mobile
phone market grew 11.3 percent in the fourth quarter of
2009, according to IDC. And the research firm estimates
that the market for voice/data mobile devices (that generally
are power hogs compared to their voice-only counterparts)
grew by nearly 30 percent year over year.
For designers of such devices, a whole series of new questions
abound. Does the system correctly power up and change
power states? Does it meet performance requirements while
powering up/down its components? Does it meet power
budgets and battery life requirements?
Answering those questions with any degree of certainty
invariably hinges on verifying those areas of the design
where software and hardware interact together the most.
Often this is a confounding task that confronts designers
with a long list of seemingly contradictory requirements.
And though no one technique is right for every design
situation, we think that a good starting point is to first
model power-management functionality at RTL and then
verify the hardware and software together in an optimized
environment.
Here’s why, and how.
ANNOTATE AN EXISTING DESIGN WITH UPF
Until recently, an engineer wanting to really drill down and
look for power-related bugs in an ASIC or SoC design faced
a series of unattractive hardware simulation choices. Gate-
level verification was highly detailed and impossibly slow.
Though marginally faster, the various ways of simulating
at RTL were complicated by the need to insert additional
power management information, which required intrusive
RTL code changes.
Figure 1: Software is relevant to most power-management functions in
low-power ASIC/SoC designs (such as switching states), which is why the
premium on hardware/software co-verification is on the rise.
24 • April / May 2010 Chip Design • www.chipdesignmag.com
FIR
MW
AR
ECO
-VER
IFIC
ATIO
N
SIM
ULATIO
N What about just focusing on software simulation to verify
applications most tightly wedded to the silicon? Though
fast, this approach too often lacked most or all of the detail
needed for debugging hardware/software interactions,
where some of the thorniest low-power issues arise.
The arrival of the Unified Power Format (UPF) changed
things for the better. A TCL-based format for specifying low-
power intent throughout the design and verification flow,
UPF was designed to allow for reuse and interoperability
between different tools. For those who care about such
things, UPF 2.0 also became an industry standard with the
adoption of IEEE Std 1801™-2009 in March 2009. (Full
disclosure: Mentor Graphics chaired the IEEE standards
activity.)
Reuse of certain functional blocks or even entire designs is
among the holy grails of ASIC/SoC design, which is ever
more costly. UPF enables such reuse, providing a relatively
straightforward means to annotate old designs with new
power management features. Engineers can supplement
existing designs with power-aware features by specifying
these in a separate UPF file. Or they can experiment with
different power control schemes by simply changing this
separate file while leaving the essential design description
alone. The alternative, continuing to tweak the RTL of
every design that requires power-management features, is
tedious and error-prone.
UPF allows for more than just defining power domains,
switches and other elements of the power architecture. It can
be used to create power strategy via power state tables; to set
up and map low power design elements such as retention,
isolation and level shifters; and to match simulation and
implementation semantics.
UPF-FRIENDLY SIMULATOR SPEEDS VERIFICATION
Unlocking the value of the UPF file requires a verification
platform built to work with the standard. Mentor Graphics
Questa is one, though there are others available from
several of the larger EDA vendors. The workflow, in short,
is to first go through and verify that the RTL actually
performs correctly, and then to toggle some settings on the
simulator and run it a second time to check the power-aware
functionality described in the UPF file. No recompilation
of the RTL is required, a boon given ever increasing gate
counts.
For all the benefits of UPF, in the end the standard is mostly
aimed at hardware verification. Advanced verification of
an ASIC or SoC loaded with power-aware features means
taking a hard look at software, or more specifically, verifying
the hardware and software together.
One way to do this is to execute the software on top of an
HDL-simulated CPU. Despite all the theoretical advantages
of tying code directly to the hardware description instead of
a higher level model, this approach can be both painfully
slow and relatively opaque. As deadlines loom and managers
turn up the pressure, too often all the engineer can say with
any authority is that something is not quite right between
the software and the underlying HDL.
Tools that speed up the simulated CPU can help matters.
Mentor’s solution, for example, replaces the HDL-based
CPU simulation with a model that’s tied in with the rest
of the logic simulation – and that operates at dramatically
higher speeds, a benefit that flows from a host of features in
the tool, including optimized memory access.
A quick primer on optimized memory access: During
verification, it’s important to first confirm that fetching
instructions from memory does in fact work. But once this
is verified, huge efficiencies are gained by abstracting it
away. In general, the more a tool avoids spending time at a
pin-wiggle-level of detail continuously checking something
that you already know works, the better.
A high-speed processor model allows design teams to run,
for example, tightly embedded RTOSs, whose importance is
rising in lock-step with the increasing need for fine-grained
management of underlying hardware. Combined with a
solid software debugger, running an RTOS can be useful in
a host of verification and debugging scenarios.
For example, imagine an engineer working on software that
controls power states. He wants to boot it up, get to a simple
prompt, and then use the software debugger to observe the
state change from turbo mode to sleep mode. The engineer
enters a command line prompt, which runs the software,
and then sits back to watch all the changes going on while
Chip Design • www.chipdesignmag.com April / May 2010 • 25
FIR
MW
AR
ECO
-VER
IFIC
ATIO
N
SIM
ULATIO
N
that software is running. One hallmark of a good tool/
verification environment is providing an engineer with pin
level visibility to both the hardware and software, or more
precisely, with an ability to closely observe when the power
control module writes out to one of the power islands and
changes its power state. Another is allowing the user to be
able to dynamically select which memory accesses run in the
logic simulator.
The speedup can be dramatic. Last fall at ARM techcon3
we presented a case where a high-speed simulator increased
the speed of embedded software execution by a factor of
10,000.
To be sure, there are alternatives to simulation-based
hardware/software co-verification. Emulation is one, a
method which can provide closer-to-final-product speeds
but often fails to provide sufficient visibility. Other emulation
headaches include increased setup time (emulation is post-
synthesis) and complexity surrounding place and route.
The real selling point of simulation is that design teams can
start doing power-related hardware/software co-verification
before their designs are done. Of course, it’s always possible
to wait and throw more people at a design or verification
problem. But in IC design, as is true throughout engineering
and most other fields as well, the earlier you can address
problems, the better.
In other words, it’s best to avoid making the call to those
crack firmware coders if you can.
* Apologies for the anonymity. But everyone knows that despite its sprawling
size (0.5% of worldwide GDP, according to Wikipedia) the semiconductor
industry is more like a small village that prizes discretion than a mega-city
that celebrates the broadcasting of every foible and failing.
Marc Bryan has been both a leading and
contributing member of tool development teams
for more than 24 years. Currently serving as
the Product Marketing Manager for Mentor
Graphics’ Codelink products, Bryan came to
Mentor after five and a half years with ARM's
tool division. A prior hands-on role at Korg R&D
provided extensive embedded processor-based, system-level design
and implementation experience.
Barry Pangrle is a Solutions Architect for Low
Power in the Engineered Solutions Group at
Mentor Graphics Corporation. He has been a
faculty member at UC Santa Barbara and Penn
State University where he taught and performed
research in high-level design automation. He has
published over 25 reviewed works in high level
design automation and low power design.
Why Software MattersBy Ed SperlingSoftware and hardware may not mix easily, and engineers on
each side of the wall may not talk the same language, but these
days no one has the luxury of ignoring one side or the other.
That message came through loud and clear at a panel discus-
sion sponsored by the EDA Consortium yesterday evening,
which included top engineers at Wind River, Green Hills and
MontaVista. Among the key facts in the discussion:
1. The majority of engineers working on an SoC are software engineers, who represent the biggest portion of the non-recurring engineering expenses.
2. A couple decades ago a typical chip had thousands of lines of embedded code. Now there are millions of lines of code, and no one person understands all of it. The
result is more complexity and a higher risk of failure—particularly when it’s not well tested with the hardware.
3. All of the major embedded software companies
except one have been bought by large semiconductor companies, which increasingly are required to include software stacks with their chips to create complete platforms for applications.
Driving these changes are some fundamental shifts in the
hardware. Jack Greenbaum, director of engineering at Green
Hills, said the shift from 8-bit bare-metal software to 32-bit
microcontrollers has opened up a huge opportunity for more
complex software. In addition, the shift from 32- to 64-bit has
allowed small devices such as microcontrollers to now start
using full-featured operating systems such as Linux because
memory is so cheap.
To read more, please visit the System-Level Design commu-
nity at: sldcommunity.com
26 • April / May 2010 Chip Design • www.chipdesignmag.com
SYSTEM
-LE
VEL D
ESIG
NCO
NSTR
AIN
T M
ETH
OD
OLG
Y
By Steve Svoboda, Cadence Design Systems
Low-Power Tradeoffs Start at the System LevelTo succeed at the end, your design methodology must weigh
all design constraints from the very beginning.
Until the early 1990s, increases in the performance of
electronics came largely without major increases in
power consumption. Starting in the late 1990s, however,
this began to change. Today, advanced central processing
units (CPUs) consume over 100 W--almost 50X the
consumption of early processors, such as the Intel i186.
Yet during this lengthy span of time, battery capacity hasn’t
been able to keep pace. Compared to the early 1980s (on a
weight-adjusted basis), for example, today’s batteries hold
only 3X to 4X as much charge. This issue poses a clear
challenge for the development of energy-efficient designs
going forward.
The primary reason for making power tradeoffs at the
system level is that the decisions made at that stage have
the greatest impact. At the system-architecture level, on
the other hand, architects select which algorithms to use,
decide which functions will be performed in software
running on processors versus custom hardware, and what
power modes the system will have. Decisions made at this
stage can easily affect system power consumption by a
factor of 10X or more.
At the next stage—the micro-architecture level—
hardware designers decide how to minimize hardware
power consumption under certain throughput and
latency constraints (i.e., what hardware operations will be
performed serially versus in parallel, how many pipeline
stages, etc.). Decisions at this stage can affect power
consumption by a factor of 2X to 10X. Beyond these two
stages, design decisions have progressively less impact on
power consumption. They also become increasingly difficult
and costly to reverse (see the Table).
LOW POWER AT THE SYSTEM LEVEL
To be effective at optimizing systems for low power, a low-
power design flow demands four capabilities:
1. Requirements capture at the specification/system
level: The designer first needs a way to capture chip-
level area, timing, power requirements, and power
budgets at the specification level. He or she must then
allocate them to different blocks within the design.
2. TLM-to-gates synthesis engine: The designer can enter
the IP portion of the design into the chip planning
tool/process. But how does he or she optimize the
blocks that were created from scratch? One way is
to create them manually at the register transfer level
(RTL) and then synthesize and estimate the power,
iterate/refine, etc.
However, a better way is to use a high-level-synthesis (HLS)
engine that enables one to do the following: create an IP
block at a much higher level of abstraction, automatically
generate and analyze different micro-architectures, and
select the one that best fits the application. The HLS
engine must be tightly integrated with the implementation
flow so that the area, timing, and power estimates that it
uses are sufficiently accurate.Table: Here, system-on-a-chip design levels are broken down according
to the impact of various decisions on overall power consumption.
Chip Design • www.chipdesignmag.com April / May 2010 • 27
SYSTEM
-LE
VEL D
ESIG
NCO
NSTR
AIN
T M
ETH
OD
OLG
Y
Once the designer has the optimized RTL micro-
architecture, he or she can make additional improvements
in power consumption using gate/implementation-level
power-saving techniques. Examples include clock gating,
multiple voltage domains with level shifters, dynamic
voltage-frequency scaling, etc.
3. Power analysis engine: A critical third element, an
analysis engine, is required to simulate the design under
“real-world” conditions. This requires a methodology
for emulating the system-on-a-chip (SoC) hardware
performance at high speed together with a mechanism
to track the toggling of the gates in the design. With
a carefully calibrated way to use the toggling, it also
must be possible to accurately estimate how much
power the gates will actually consume.
4. Power-aware verification methodology: Power-saving
techniques like those mentioned above (clock gating,
multiple voltage domains, dynamic voltage frequency
scaling, etc.) significantly affect the functional behavior
of the SoC. They also can add major complexity to the
verification process, unless one has the proper tools
and methodology to deal with these issues.
Tying these four elements together lets designers deal
with the challenges of lower-power SoC design in a timely,
cost-effective way with higher productivity.
A POWER-AWARE WORKFLOW
Companies have begun working to address the requirements
of low-power SoC design at the system level. For
requirements capture, for example, a product called InCyte
Chip Planning promises to become the “cockpit” where the
high-level decisions and constraints on individual blocks
are made. As the different design teams implement blocks
within the system and determine whether they’ve met
unit-level specifications, this product allows them to feed
that information back to chip planning. This collaborative
process enables system-level refinements to be made. If one
block team is able to do better than their specifications,
“slack” may be freed up for another block team, which may
be struggling to meet their specifications.
For hardware synthesis, a product called C-to-Silicon
Compiler combines high-level synthesis with conventional
logic synthesis. It vows to let designers automatically convert
TLM system-level models into RTL and then to gates while
optimizing for area, timing, and power. By having the high-
level and logic synthesis integrated together in this fashion,
the compiler ensures that whatever RTL micro-architectures
are generated will be physically implementable.
Although the static timing and power analysis done by the
synthesis tools is indispensable to ensure correct hardware
implementations, it is not sufficient. Because actual
peak/average power consumption can vary tremendously
with operating conditions, dynamic analysis capability is
required as well. For this purpose, the Palladium product
family promises to estimate SoC power consumption
under “real-world” conditions while actual system software
is being executed.
Finally, verification consumes 60% to 70% of the R&D
effort on today’s SoC-development projects. With power-
aware designs, verification challenges grow more intense
as special register-transfer/gate-level features to reduce
power (e.g., clock gating) add further circuit complexity.
Power-management features are added that are under
hardware and software control. In addition, various design
modes must be verified at different operating voltages. The
net result is a huge expansion of the system state space,
which must be verified. Here, power modes become yet
another dimension of design parameters. All (or as many
as possible) combinations of power and operating modes
must be verified or else problems will arise (such as data
loss, dead-lock conditions, etc.).
Power-aware, metric-driven verification with technologies
like Conformal and Incisive goes a long way toward solving
these challenges. With these technologies, power-intent
information of the design is captured (in a format like CPF).
It is then used during verification to ensure that the design
behaves as it would with all of the power-control logic in
the RTL. These tools are used to infer all of the power
modes that need to be covered in the design, automatically
creating the appropriate coverage metrics and assertions.
Verification then continues as normal. Now, however, it is
fully power-aware.
28 • April / May 2010 Chip Design • www.chipdesignmag.com
SYSTEM
-LE
VEL D
ESIG
NCO
NSTR
AIN
T M
ETH
OD
OLG
Y
THE OLD VERSUS THE NEW
In the “classical” flow, teams are naturally reluctant to refine
their RTL for a 20% to 30% QoR gain (see the Figure).
This is because design-iteration loops take too long to be
useful for a project that typically lasts six to nine months. In
the new flow, however, greater automation and integration
allow iterative steps to be completed much faster and—in
many cases—in parallel. With a modern HLS engine,
logic-synthesis comes built-in. This capability enables
the designer to go directly from TLM SystemC to gates.
There’s no longer any need to fine-tune RTL or create
synthesis scripts, as must be done in the manual approach.
Similarly, when the logic-synthesis engine is integrated
with a hardware-emulation system (as in the case of RTL
Compiler with Palladium), calibration between toggle-
counts obtained from the emulation system and the actual
power consumption represented by those toggles at the gate
level is already done. It is guaranteed to be accurate as well.
With the major portion of today’s electronic products
targeting mobile applications, power consumption has
evolved to become a primary design constraint. Any effective
design flow and methodology must simultaneously consider
all design constraints (including power) in a seamless closed-
loop, multi-objective, planning-to-signoff solution.
The major recent EDA industry achievement has been to
extend this flow upwards in abstraction from RTL to TLM to
include power exploration, estimation, and analysis at every
step including SystemC/TLM design exploration, software
optimization, hardware (TLM and RTL) synthesis, and
physical design/signoff. In parallel with the implementation
flow, power verification has been extended. By leveraging
static, dynamic, and formal power-verification techniques
in a closed-loop verification methodology, design teams can
avoid last-minute power-related surprises. The net result is
to enable first-pass silicon success with 5X to 10X higher
engineering productivity.
Steve Svoboda is Technical Marketing Di-
rector for System Design and Verification
solutions at Cadence Design Systems. He
is a 15-year EDA veteran with Master's
degrees in both Electrical Engineering and
Engineering Management from Stanford
University, as well as undergraduate degrees
in EE and Economics from Johns Hopkins.
Figure: Here, the classical RTL and new TLM-driven design flows are
compared.
A Positive Indicator of Semiconductor DirectionBy Jim LipmanI think the San Jose version of the annual TSMC Technology
Symposium this past week is a good indicator of where the
semiconductor industry is going over the next few years. The
positive growth predictions of keynote speaker Morris Chang,
TSMC founder, chairman and CEO - 22% this year and 7%
in 2011 - are only one gauge of industry direction. Another is
what happens on the exhibit floor.
As in past years, there were plenty of companies willing to put
out the money and resources for an exhibit floor booth in San
Jose (which also includes smaller, tabletop exhibits this month
in Austin and Boston). From a potential customer perspec-
tive, foundry-driven shows such as the TSMC symposiums
provide an excellent opportunity to meet prospects who
are, largely, in the chip design business. Since my company,
Sidense, is an IP provider, this is a good audience for us.
However, the number of exhibitors is not a great measure of
industry direction - the attendee base serves this purpose.
To read more, please visit "Turning to Jim" blog at: http://
www.chipdesignmag.com/lipman/
Chip Design • www.chipdesignmag.com April / May 2010 • 29
ALG
OR
ITH
M O
PTIM
IZATIO
NLO
W P
OW
ER
Take A New Approach to the Power-
Optimization of Algorithms and Functions
The power consumption of digital integrated circuits (ICs)
has moved to the forefront of design and verification
concerns. In the case of handheld, battery-powered devices
like cell phones, personal digital assistants (PDAs), e-books,
and similar products, users require each new generation to
be physically smaller and lighter than its predecessors. At the
same time, they expect increased functionality and demand
longer battery life. It’s therefore obvious why low-power design
is important in the context of this class of products. In reality,
however, low-power considerations impact almost every modern
electronic system—including those powered from an external
supply.
CHALLENGING THE CONVENTIONAL WISDOM
The architecture of a system is a first-order determinant of that
system’s power consumption. When it comes to the functional
blocks themselves, the hardware design engineer must determine
the optimal micro-architecture for each block. Different micro-
architectures have very different area, timing, latency, and power
characteristics. The register transfer level (RTL) is the earliest
stage of design abstraction at which it’s possible to gain sufficiently
accurate estimations of characteristics like area and power. If
created by hand, however, RTL is very fragile, complex, and
time consuming to capture. As a result, there’s typically sufficient
time to create only one micro-architecture (or a very limited
number of micro-architectures). A wide range of alternative
implementation scenarios therefore remains unexplored. In
addition, RTL does not support sophisticated parameterization,
so an IP block cannot be retargeted into multiple systems-on-a-
chip (SoCs) with different area/speed/power targets.
The ideal scenario is to have an environment in which design
engineers can create and functionally verify behavioral
representations at a high level of abstraction. They should then
be able to quickly and easily convert these representations into
equivalent RTL for detailed power analysis. Furthermore, this
ideal scenario includes the ability to be able to create a single
behavioral representation and to use it to generate and evaluate
a full range of alternative RTL implementations. Currently, the
predominant high-level alternative to RTL-based design has
been to use sequential programming-based C/C++/SystemC
representations in conjunction with some form of behavioral
synthesis. While these approaches can raise the design’s level of
abstraction, they have significant limitations. For example, they
provide poor quality-of-synthesis results except for the narrow
range of application spaces that they can efficiently address.
Additional issues include the following:
• The model of computation in C/C++/SystemC,
sequential, threaded, flat memory has been fine-tuned to
execute on von Neumann computing platforms. This approach
is inappropriate for hardware designs that feature fine-grain
parallelism and heterogeneous storage. As a result, common C/
C++ idioms and style (loops, pointers, byte-oriented data types,
etc.) must be laboriously “rewritten” to prepare an application for
C/C++ synthesis.
• C/C++ synthesis tools work by customizing a few generic
template architectures. As a result, there’s not much room
for architectural variation. Achieving good quality often
requires a “long tail” of effort, massaging both source code
and constraints in tool-specific ways because of the non-
transparent effect on architecture. The resulting code/
constraints aren’t portable or maintainable. In addition,
the ad-hoc, proprietary constraint languages may not be
parameterizable, requiring multiple sets of source code and/
or constraints to cover the required architectural space.
• Automatic parallelization of sequential code is feasible
mainly for digital-signal-processor (DSP) -like (“loop and
array”) applications. This restricts its use to only a few blocks
of the design. Even within this “sweet spot,” it’s difficult to
address essential “system issues,” such as memory sharing,
caching, pre-fetching, non-uniform access, concurrency, and
integration into the full chip design.
• Loss of control with regard to the process of timing
closure also is a problem. When compiling a “behavioral”
C/C++ description into hardware, the semantic model
of the source (the sequential code) is so different from the
semantic model of the ensuing hardware that the designer
loses predictability. It’s difficult for the designer to imagine
what should be changed in the source to effect a particular
desired improvement in the hardware. Furthermore, small
and apparently similar changes to the source can result in
radically different hardware realizations.
By Rishiyur S. Nikhil, Bluespec, Inc.
30 • April / May 2010 Chip Design • www.chipdesignmag.com
LO
W P
OW
ER
ALG
OR
ITH
M O
PTIM
IZATIO
N
The ability to evaluate a wide range of micro-architectures can
result in more optimal results than painstakingly hand-coded
RTL. For many of the reasons listed above, however, not all
high-level (behavioral) languages and associated HLS engines
facilitate the ability to automatically generate the full range of
micro-architectures for evaluation.
In the same way that it would be unimaginable for software
developers to neglect to evaluate alternative algorithms, it should
be unimaginable for hardware designers to proceed without
considering alternative micro-architectures. In reality, however,
the lack of rigorous micro-architecture evaluation is the norm
rather than the exception. But what if design engineers had the
ability to quickly and easily evaluate the entire gamut of micro-
architecture alternatives ranging from highly parallel to highly
serial implementations?
A NOVEL APPROACH
A new approach, such as PAClib, has emerged using a plug-and-
play library of common pipeline building blocks. Those building
blocks are designed for constructing algorithms and datapath
functions as illustrated in Figure 1. (This is, of course, a very
simple representation; PAClib modules can be instantiated by
other modules and wrappers and so forth.)
For example, suppose a PAClib pipeline module called PF
computes some function f(x) on each input x. In addition, a
pipeline module called PG computes some function g(y) on
each input y. By cascading PF followed by PG, it’s possible to
construct a pipeline that computes the function g(f(x)) on each
input x. Alternatively, by cascading PG followed by PF, designers
can construct a pipeline that computes the function f(g(y)) on
each input y.
These PAClib building blocks are highly parameterized.
They allow designers to separately specify attributes like
computational functions, pipeline buffer insertion, and
pipeline structure choices—all without having to worry about
actual implementation details. By simply “dialing in” different
parameters, alternative micro-architectural versions are
automatically generated with correct pipeline control logic and
implementation details.
This PAClib library is written in Bluespec SystemVerilog
(BSV). It augments standard SystemVerilog with rules and
rules-based interfaces that support complex concurrency and
control across multiple shared resources and across modules.
BSV features the following: high-level abstract types; powerful
parameterization, static checking, and static elaboration; and
advanced clock specification and management facilities. One
of the key advantages is that the semantic model of the source
(guarded atomic state transitions) maps very naturally into
the semantic model of clocked synchronous hardware. BSV’s
computation model is universal (equally suitable for datapath
and control). So it can directly address system considerations,
such as memory sharing, caching, pre-fetching, non-uniform
access, concurrency, and integration into the full chip design.
With full architectural transparency, the designer also can
make controlled changes to the source with predictable effects
on timing. Due to the extensive static checking in BSV, these
changes can be more dramatic than the localized “tweaking”
techniques favored when working with standard RTL. As
a result, designers can achieve timing goals sooner without
compromising correctness. The end result of using PAClib is
that hardware design engineers continue to think like design
engineers. But they now have access to rapid algorithmic design
and architectural exploration capabilities.
24 MICRO-ARCHITECTURES FROM A SINGLE SOURCE
As an example, look at the Inverse Fast Fourier Transform
(IFFT) block used in the IEEE 802.11a transmitter system (see
Figure 2). The term 802.11a refers to a common IEEE standard
for wireless communication. The protocol translates raw bits
from the media access controller (MAC) into orthogonal-
frequency-division-multiplexing (OFDM) symbols in the form
of sets of 64 32-bit, fixed-width, complex numbers. The protocol
is designed to operate at different data rates. At higher rates, it
consumes more input to produce each symbol. Regardless of
the rate, all implementations must be capable of generating an
OFDM symbol every 4 ms.
For the purposes of these discussions, it’s not necessary to
understand the 802.11a transmitter in any level of detail. It’s
sufficient only to appreciate that the IFFT block can account
for approximately 90% of the silicon real estate, depending on
its implementation. Additionally, the critical timing path of the
IFFT is many times larger than the critical path of any other
block in the system. Thus, the focus here is on this block.
Figure 1: This graphic shows how PAClib module pipeline interfaces plug
together.
Chip Design • www.chipdesignmag.com April / May 2010 • 31
ALG
OR
ITH
M O
PTIM
IZATIO
NLO
W P
OW
ER
The IFFT is constructed from two basic computational functions,
f_radix4 and f_permute, which are treated here as black boxes.
Conceptually, the IFFT is a cascade of three identical stages as
illustrated in Figure 2. The input and output of each stage—and
of the IFFT as a whole—are vectors of 64 complex numbers
with 16-bit real and imaginary parts. Each stage also receives
a set of coefficients, which may be different for each f_radix4
instantiation.
The 64-element input vector to each stage is partitioned into 16
slices—each comprising four complex numbers. Each group of
four complex numbers is fed into an f_radix4 function, which
also has four complex number outputs. Thus, the outputs from a
column of 16 f_radix4 functions acting in parallel also are a 64-
element vector of complex numbers. This vector is fed into an
f_permute function, which permutes the vector. It then outputs
another 64-element vector of complex numbers, which forms
the output from this stage.
The mathematical details of the IFFT aren’t important for the
purposes of this article. Instead, the focus should be on the
underlying structure of the computation (see Figure 3). The
goal is to investigate how alternative micro-architectures—all of
which compute the same mathematical function—may differ in
area, performance (throughput, clock speed, latency, etc.), and
power.
One possibility would be to implement the IFFT as a purely
combinatorial circuit (see Figure 4a). Another alternative would
be to add pipeline buffers to the outputs of the f_permute
functions as illustrated in Figure 4b and also to the inputs of the
f_permute functions as illustrated in Figure 4c. These buffers
increase the hardware cost. Yet they will likely decrease the
critical path length and allow synthesis at a higher frequency,
thereby increasing overall throughput. Yet another possibility is
to implement just one stage, but to loop the data through this
stage three times to replicate the actions of the three stages (see
Figure 4d).
For any of the preceding choices (except the purely combinatorial
implementation), it also is possible to vary the micro-architecture
implementations of each “stack” of f_radix4 functions. Instead of
16 f_radix4 functions, the designer might “funnel” or “serialize”
the 64-element input vector into two 32-element vectors. He
or she can run each of these vectors through the same group
of 8 f_radix4 instances and then “unfunnel” or “deserialize” the
emerging sequence of two 32-element vectors back into a 64-
element vector.
Alternatively, the designer could funnel the 64-element input
vector into four 16-element vectors and run each of these
vectors through the same group of four f_radix4 instances.
Or the 64-element input vector can be funneled into eight 8-
element vectors. Each of these vectors can be run through the
same group of two f_radix4 instances (see Figure 5).Yet another
alternative is to funnel the 64-element input vector into 16 4-
element vectors and run each of these vectors through a single
f_radix4 instance.
Figure 2: This high-level block diagram depicts an IEEE 802.11a
transmitter.
Figure 3: This high-level dataflow graph details the IFFT block.
Figure 4: Some alternative micro-architectures for the IFFT block are
shown here.
32 • April / May 2010 Chip Design • www.chipdesignmag.com
LO
W P
OW
ER
ALG
OR
ITH
M O
PTIM
IZATIO
N
Because they re-use hardware, looping and funneling may seem
to be guaranteed to reduce silicon area. Yet these techniques
require additional buffers and control circuitry. Furthermore, for
a given target throughput, they will require higher clock speeds
with an associated cost in power consumption. The end result is
that there are a wide variety of potential micro-architectures. It
is difficult to predict which one will be “best” for a given set of
throughput, area, and power goals.
The PACLib library’s family of plug-and-play building blocks
allows all of the micro-architectures envisioned for the IFFT in
Figure 4 to be expressed in a single source using only 100 lines of
code. It is parameterized by the f_radix4 and f_permute modules
and the micro-architectural choices. Simply by controlling the
parameters associated with this design, this single high-level
representation can be quickly and easily used to generate, in this
example, 24 different micro-architectures. The design engineer
can then determine which structure provides the optimum
combination of throughput, area, and power characteristics
to address the requirements of his or her particular target
application.
In closing, the predominant high-level alternative to RTL-
based design has been to use sequential programming-based,
C/C++/SystemC representations in conjunction with some
form of behavioral synthesis. These approaches can raise the
design’s level of abstraction. Yet they have significant limitations
including poor quality-of-synthesis results except for the narrow
range of application spaces that they can efficiently address. In
addition, they lack the parameterization needed to be able to
express multiple micro-architectures uniformly. A new approach
can be found in solutions like PAClib. This plug-and-play library
of common pipeline building blocks is designed for constructing
algorithms and datapath functions. The PAClib library is
written in Bluespec SystemVerilog (BSV), which augments
standard SystemVerilog with rules and rules-based interfaces
that support complex concurrency and control across multiple
shared resources and modules. BSV’s computation model is
equally suitable for datapath and control.
By capturing an algorithmic design using PAClib and then
controlling the parameters associated with the PAClib modules,
a single high-level representation can be quickly and easily used
to generate multiple alternative micro-architectures. As a result,
design engineers can focus on determining which structure
provides the optimum combination of throughput, area, and
power characteristics to address the requirements of their
particular target applications. The end result is to dramatically
reduce the development cycle while significantly improving the
quality of results. Such approaches are predictable and seamlessly
integrate complex control. As a result, they don’t suffer from the
“long tail” of effort and difficulty addressing essential “system
issues,” such as memory sharing, caching, pre-fetching, non-
uniform access, concurrency, and integration into the full chip
design.
Rishiyur S. Nikhil is co-founder and
CTO of Bluespec Inc. Previously, he led
a team inside Sandburst Corp. that was
developing Bluespec technology. Nikhil
also served as acting director at Cambridge
Research Laboratory (DEC/Compaq)
and was a professor of computer science
and engineering at MIT. He holds patents
in functional programming, dataflow and multithreaded
architectures, parallel processing, compiling, and EDA. Nikhil
received his PhD and MSEE in computer and information
sciences from the University of Pennsylvania. He received
his bachelors in technology in electrical engineering from IIT
Kanpur.
Figure 5: Here, the micro-architecture of the f_radix4 “stack” is varied.
Chip Design • www.chipdesignmag.com April / May 2010 • 33
DOT.ORG
I ’ve served as president of the IEEE Council on EDA
(also known as CEDA), a focal point since 2005 for EDA
activities spread across six IEEE societies — Antennas and
Propagation; Circuits and Systems; Computer; Electron
Devices; Microwave Theory and Techniques; and Solid
State Circuits.
Since its formation, CEDA has worked to expand its
support of emerging areas within EDA and brought more
recognition to members of the EDA profession. And as
my two-year term ends, I look back with pride at the way
in which CEDA’s Executive Committee has found many
creative ways to further expand support of the EDA
community that makes me quite proud.
Just announced is the formation of the Design Technology
Committee, a group of executives from EDA user
companies. DTC has the goal to work with groups inside
and outside the IEEE to promote best practice sharing
and strategic solutions to address gaps between EDA
capabilities and future needs.
CEDA now sponsors 14 EDA conferences and workshops,
including DAC, ICCAD, DATE and ASPDAC. It created
the Embedded Systems Letters, a new publication for
rapid communication of short notes in an increasingly
important EDA area. This adds to the quarterly Currents
newsletter and the mainstay TCAD Journal.
Two new awards presented in 2009 help to recognize the
accomplishments of members of our community. The
yearly Richard Newton Technical Impact Award is jointly
sponsored with the ACM Special Interest Group on
Design Automation (ACM SIGDA). It is awarded to an
individual or individuals for their outstanding technical
contributions to EDA, recognized over a significant period
of time. The Early Career Award, also to be presented
yearly, recognizes an individual who has made innovative
and substantial technical contributions to EDA in the
early stages of their career
These awards complement the existing awards:
• The D. O. Pederson Best Paper Award presented in
the TCAD Journal
• The William McCalla Best Paper Award presented at
ICCAD
• The prestigious Phil Kaufman Award jointly sponsored
with the EDA Consortium
The ongoing Distinguished Speaker Series offers a number
of complementary events at EDA conferences and will
continue in 2010. CEDA has also begun experimenting
with new ways of reaching the EDA community, including
a live webcast of the ICCAD keynote. A digital edition of
Design and Test Magazine is available for a reduced fee and
the first issue of Embedded Systems Letters is available
online, while an on-line calendar of EDA key conference
dates can be found at www.ceda-org.
Cadence’s Andreas Kuhlmann will become CEDA’s
president in January and is committed to continued support
of the EDA community with more valuable activities in
2010. He and the rest of CEDA’s Executive Committee
would benefit by having your help. I encourage you to
visit the CEDA website (www.c-eda.org) to learn more
about us and to get more involved. Organizations such as
CEDA are driven by the efforts of volunteers.
John Darringer is President of the IEEE
Council on EDA. He can be reached at
[email protected]. For more information
on CEDA visit http://www.c-eda.org
Creatively Supporting the EDA Community
By John Darringer, President of the IEEE Council on EDA
Register today at the Early Bird rates & save up to $200 on yourConference Pass! Or, register now for a FREE Expo Hall Pass!
Co-located with the Embedded Systems Conference Chicago!
Find the Solutions to Your Sensors &Sensing Technology Challenges!
Subscribers: Visit: www.sensorsexpo.com to register or call 877-232-0132 or 972-620-3036 (outside U.S.). Use discount code
F327M for an EXTRA $50 OFF a Gold or Main Conference Pass!
• Energy Harvesting• Low-Power Sensing• Wireless Networking• Bio-Sensing
Gain the knowledge you need from leading experts and peers in the sensors industry.
• MEMS & MCUs• Monitoring Tools & Applications• Novel Approaches to Measurement• Power/Smart Grid Monitoring & Control
This year’s Conference Program includes more than 40 Technical Sessions in 8 Tracks covering:
Identify specific solutions to your most difficult detection and control-related challenges on the expo floor.
Sensors Expo brings together the largest and best-in-class showcase of sensing technologies and systems for attendees to evaluate and make informed decisions.
PRODUCED BY: OFFICIAL PUBLICATION: SILVER SPONSOR: MEDIA SPONSOR:
Conference: June 7-9, 2010 • Exhibits: June 8-9, 2010Donald E. Stephens Convention Center • Rosemont, IL • www.sensorsexpo.com
Chip Design • www.chipdesignmag.com April / May 2010 • 35
TOP VIEW
By K. Charles Janac, Arteris Holdings
Today’s SoC designers are at an inflection point that requires
a re-thinking of how complex designs are implemented.
They are challenged with the increasing amount of IP blocks
that must be managed on a single chip, a trend that introduces
complexity, performance and power issues which strain
traditional approaches to implementing an SoC ‘infrastructure.”
As a result, Network-On-Chip (“NoC”) technology is rapidly
displacing traditional bus and crossbar approaches for SoC on-
chip interconnect. The reason is simple: NoC architectures result in
smaller, faster and lower power on-chip-interconnects than previous
approaches. With viable commercial offerings that have been proven
in productions now available on the market, NoC’s time has arrived
as a mature and efficient way to help SoCs scale.
High performance requirements, quality of service (QoS) needs,
and physical design constraints make the integration of an increasing
number of heterogeneous IP cores in a SoC a formidable challenge.
It’s made even more challenging because traditional on-chip
interconnect is taking up an increasingly large part of the SoC design,
not only in chip area, but in design complexity and design resources
as well. Previous generation approaches such as a bus, crossbar or
hybrid combinations of buses and crossbars cannot meet area and
power budgets, or frequency targets. Even in the smaller designs,
power requirements alone make traditional interconnect approaches
sub-optimal.
Designers are moving toward NoC architectures because they offer a
light-weight packet-based communication system that meet stringent
area and power requirements, without sacrifices in performance.
In a NoC, data-packets travel between processing elements in the
interconnect through physical links, and where the physical properties
of each link can be independently configured according to bandwidth,
latency, clocking, power or physical design requirements. Interconnect
processing elements can be switches, data-width converters, clock-
domain converters, power-isolator blocks, security modules and
others. Packet-based formatting allows the processing elements to be
very simple and link configurability provides the optimal link design
for each communication path.
A common misconception is that NoC architectures introduce
additional latency. But robust NoC solutions allow for packet
configurability. A flexible packet format allows the designer to make
the optimal wiring versus latency trade-off on a connection-by-
connection basis, including the configuration of packet formats that
do not insert any additional latency as may be required by the CPU
to DRAM path.
The flexible topology enabled by the true NoC approach allows
designers to independently address the many design constraints
imposed upon the interconnect by the SoC. The NoC architecture
removes the need to divide up the SoC into independent islands,
each governed by a standalone arbitration scheme, and requiring all
interconnected IP to conform to the same interface standard, clock
and data width. The ability of the NoC architecture to address
unrelated constraints in an orthogonal way leads to superior results
and improved SoC economics.
Because of the complexity and design requirements of modern SoCs,
it is inevitable that a NoC approach will become the de fact way to
develop these devices.. The question remains whether semiconductor
companies will develop such technology in-house or will opt to license
it. Over the last fifteen years, the industry has been relatively slow in
moving towards licensing interconnect technology. The reasons have
been twofold: the existence of prior infrastructures and not enough
“bang-for-the-buck” in commercial interconnect products to warrant
the switching cost. Now, however, commercial NoC products are
outperforming traditional internal solutions by significant margins,
and the costs faced by semiconductor vendors for the internal
development of a new interconnect infrastructure around an NoC
architecture cannot be justified. Indeed, the semiconductor vendors
that are using NoC technology for on-chip interconnect today are
already reaping the benefits of bringing faster, lower power and
cheaper products to the market.
K. Charles Janac is the Chairman, President
and Chief Executive Officer of Arteris Holdings.
Charlie has over 20 years experience building
technology companies. He was employee number
two of Cadence Design Systems, and later served
as CEO of HLD Systems, Smart Machines and
Nanomix. Born in Prague, Czech Republic, he
holds both a B.S. and M.S. degrees in Organic Chemistry from Tufts
University and an MBA from Stanford Graduate School of Business.
He holds a patent in polymer film technology.
NoC Technology Offers Smaller, Faster,
and More Efficient Solutions
36 • April / May 2010 Chip Design • www.chipdesignmag.com
By Dr. Joost van Kuijk, CoventorNO RESPINS
Microelectromechanical systems (MEMS) are micro-
scale or nano-scale devices. Typically, they’re fabricated
in a manner similar to integrated circuits (ICs) to exploit the
miniaturization, integration, and batch processing benefits of
semiconductor manufacturing. Yet unlike ICs, which consist
solely of electrical components, MEMS devices combine
technologies from multiple physical domains. They may contain
electrical, mechanical, optical, or fluidic components.
Spurred by growth in consumer electronics, the total market for
MEMS is projected to grow more than 40 percent from 2008 to
2012. It will go from just over $7 billion worldwide to over $13
billion, according to market research firm Yole Development. The
market for MEMS in mobile phones is expected to grow by more
than 4X during this period to $2 billion.
As promising as these forecasts sound, only a few large IDMs are
well positioned to benefit from this rapidly growing market. This is
due to the specialized expertise, long development time, and high
cost of bringing MEMS devices to market. Almost all MEMS
devices are tightly integrated with electronics—either on a common
silicon substrate or in the same package. Yet MEMS design has
traditionally been separated from IC design and verification.
MEMS devices are typically designed by PhD-level experts in
such fields as mechanical, optical, and fluidic engineering. They
use their own two-dimensional (2D) and three-dimensional (3D),
mechanical computer-aided-design (CAD) tools for design entry
and finite-element-analysis (FEA) tools for simulation. Eventually,
the MEMS design must be handed off to an IC design team in
order to go to fabrication. But the handoff typically follows an
ad-hoc approach that requires a lot of design re-entry and expert
handcrafting of behavioral models for functional verification.
Moreover, MEMS historically requires specialized process
development for each design, resulting in a situation often
described as “one process, one product.” While there are a number
of specialized MEMS foundries, support from pure-play foundries
has been very limited. According to one analyst report, it takes an
average of four years of development and $45 million in investment
to bring a MEMS product to market.
Several trends are converging to make this level of effort and
expertise unacceptable—not only for new entrants in the MEMS
market, but for the best-positioned IDMs as well. First, the fast-
paced consumer-electronics market demands design cycles that
are measured in months, not years. In addition, design costs must
be such that they can quickly maximize return on investment and
profitability.
Secondly, the market is demanding more functionality from MEMS
devices. For example, enhanced sensitivity requires that more
analog and digital circuits will be placed around MEMS devices.
The third trend is the rise of advanced packaging technologies, such
as system-in-package (SiP) and chip stacking with through-silicon
vias (3D IC). These technologies will allow manufacturers to
package all of this functionality more densely, combining multiple
MEMS sensors with analog and digital dice in a single package.
These technologies will allow manufacturers to package all of this
functionality more densely, combining multiple MEMS sensors
with analog and multiple digital die in a single package.
These demands make MEMS more susceptible to unwanted
coupling between sensing modes as well as between the MEMS
sensors and electronics. The present approach to MEMS design—
with separate design tools and ad-hoc methods for transferring
MEMS designs to IC design and verification tools—is simply not
up to these new challenges. The time has come to “democratize”
MEMS design and bring it into the IC design mainstream. The
result would be reduced design costs and shortened time to market.
In addition, the MEMS design would no longer be confined to
teams of specialists inside IDMs.
A critical key to accomplishing this “democratization” is to build
an integrated design flow for MEMS devices and the electronic
circuits with which they interact. A structured design approach
should be used that avoids manual handoffs. Companies like
Coventor and Cadence are now working together to develop such
integrated methodologies. Their goal is to shield IC designers from
the complexity of MEMS design while reducing the time, cost, and
risk of developing MEMS-enabled products.
Dr. Joost van Kuijk is vice president of marketing
and business development at Coventor. Dr. van
Kuijk has more than 16 years of experience in
the MEMS field, specializing in modeling and
simulation. He received a PhD in micro system
technology from Twente University, where he also
received a diploma in technology information.
In addition, Dr. van Kuijk holds an MSc in
mechanical and precision engineering from Delft University.
MEMS Is Poised to Cross the Chasm
The Global Semiconductor Alliance (GSA) mission is to accelerate the growth
and increase the return on invested capital of the global semiconductor industry
by fostering a more effective fabless ecosystem through collaboration,
integration and innovation. It addresses the challenges within the supply chain
including IP, EDA/design, wafer manufacturing, test and packaging to enable
industry-wide solutions. Providing a platform for meaningful global collaboration,
the Alliance identifies and articulates market opportunities, encourages and
supports entrepreneurship, and provides members with comprehensive and
unique market intelligence. Members include companies throughout the supply
chain representing 25 countries across the globe.
GSA Member Benefits Include:
Access to Profile Directories
Ecosystem Portals
IP ROI Calculator
IP Ecosystem Tool Suite
Global Semiconductor Funding, IPO and M&A Update
Global Semiconductor Financial Tracker
Semiconductor End Markets
MS/RF PDK Checklist
AMS/RF Process Checklist
MS/RF SPICE Model Checklist
Consumer Electronics Study
Discounts on various reports and publications including:
Wafer Fabrication & Back-End Pricing Report
Understanding Fabless IC Technology Book
IC Foundry Almanac, 2009 Edition
SiP Market & Patent Analysis
Global Exposure Opportunities:
Advertising
Sponsorships
Member Press Section on GSA Web site
Company Index Listing on Ecosystem Portals
Member Spotlight on GSA Web site
12400 Coit Road, Suite 650 | Dallas, TX 75251 | 888-322-5195 | T 972-866-7579 | F 972-239-2292 | www.gsaglobal.org
One system, infinite verification possibilities.
First, EVE's ZeBu emulators broke the billion-cycle barrier. Now, ZeBu-Server - EVE's next generation emulation system - has broken the billion-gate barrier. Scalable to handle up to 1 billion ASIC gates, and with execution speeds up to
30MHz, ZeBu-Server is a multi-mode, multi-user emulator suitable for all system-on-chip (SoC) hardware-assistedverification needs, across the entire development cycle, from hardware verification and hardware/software integration
to embedded software validation.
Used in virtually every ASIC/SoC industry, from graphics and computer peripheral applications, to processor and wireless mobile applications, ZeBu emulators are truly a universal solution.
ZeBu-SServveer: Billions of Cycles for Billionnss off GatesZeBu-Server: Billions of Cycles for Billions of Gates
© 2010 EVE. All rights reserved. Contact us at 408-457-3200 or 888-738-3872 (toll-free) // www.eve-team.com // [email protected]
Contact EVE-Team today for a FREE consultation at [email protected], or visit us online at www.eve-team.com.
Speed, Capacity, & Lowest Cost of Ownership.