I sat  in during the User Track session on Chip-Package-System power delivery network design optimization.

Sorin Dobre from Qualcomm presented how they use a comprehensive time and frequency domain methodology to design and validate their wireless chip-package-board designs.  He highlighted the complexities associated with  time domain analysis in which he includes the chip layout, package RLCK model and a fitted model for the PCB in a multi-domain analysis framework. He uses a multi-cycle resonance frequency aware VectorLess methodology to define the excitation model for his time domain analysis of this system. The current signature for the final excitation model generates high energy around the system resonance frequency stressing the chip-package-system PDN.

Davide Pandini from ST Microelectronics, Agrate, presented results on power supply noise and EMI analysis and how it can be mitigated for the chips his group designs for the automotive market. He presented in the second part of his talk the results on a case study highlighting static and dynamic voltage drop and EMI results. He illustrated how power noise from the chip propagates along the chip/package/board supply lines and radiates to the air. He highlighted a methodology of controlling the dynamic switching on the chip to control the on-die noise/drop and hence the EMI coupling.

Erhan Erglin from AMD presented a methodology of dynamic voltage drop analysis for their high performance designs targeting CPU/GPU/APU applications. He presented arguments on the need for dynamic voltage drop analysis and why inductance, both on the package and on the chip have to be included for their designs using advanced processes. As supply voltages reach threshold voltage levels, dynamic voltage drop analysis highlights areas of switching noise that can cause the chip to fail. Static IR which has been used historically is useful to find gross violations and typically show small drop (3-4% range) for their designs. But to include the package/die inductance and capacitance and the simultaneous switching current, he highlighted the need for dynamic voltage drop analysis. He presented findings showing that the inclusion of on-die inductance changes not only the peak voltage drop number, but also changes the voltage drop map.

Ricky Yong from Intel, Penang, Malaysia, presented results on power noise analysis on their MTCMOS (power gate)  designs including silicon correlation results. He focused on dynamic power noise analysis for mode transition for his power gated designs. He highlighted how they perform on-die voltage drop measurements using sense lines and showed close match between measurement and simulation results. He also presented simulation results for different power-on sequences and correlation to measurements (within 4%).

Kyung-Tae Do from Samsung presented a methodology for estimating statistical leakage and the library/modeling support needed to achieve that. He highlighted that present techniques are not applicable for large macros and that a new technique is needed. His proposed methodology uses a combination of library characterization using Monte-Carlo simulations and silicon measurement based tuning.

Souvik Mukherjee from TI, Dallas, showed results on a different but equally critical area of IO sub-system design and the impact of power ground noise on IO/memory interface timing. He mentioned that an ideal SSO/SSN analysis methodology needs to accurately trade off between accuracy and efficiency of modeling, extraction and simulation flows. He presented results using the Sentinel-SSO technology from Apache on how the PDN noise and signal cross-talk impact can be included for IO/memory interface timing analysis. He presented results using this technique and compared that to silicon measurement results for a 45nm SoC based wireless platform.

Again thanks to the various DAC committees and to all the folks who made DAC 2010 a very useful and productive show.  Headed back home now

At advanced nodes, process rules make full hand layout almost impossible. For example, instead of simple space rules, “space” now depends on the width of metal and the length of parallel lines, and there are complex via and contact density rules and end-of-line rules that can’t readily be dealt with by hand crafting. An automatic custom design routing tool that can deal with the new custom world will improve productivity and achieve on-time, efficient design delivery.

The thought of automation raises the specter, with some designers, that they, or the majority of their functions, will be replaced. However, increased custom design automation will increase the productivity of designers, not replace them, as we have seen in the digital design world. Given that extremely complex projects now need to be completed in the same time and with the same number of people as older, simpler designs, automating the custom design process to increase the productivity of designers is the only way to manage multi-thousand-gate designs.

The list of “must haves” for today’s routing technologies includes more flexibility, adherence to multiple process rules, and faster performance. A router should allow arbitrarily-shaped cell row regions to be generated for arbitrarily-shaped cells, and let them be continuously reshaped and developed for optimum area efficiency. It should also offer specialized routing functions for extreme-aspect-ratio designs. The routing engine needs to complement the skills and knowledge of an experienced layout designer. Interactive and semi-automatic editing features that are correct by construction should, at minimum, provide an on-line DRC capability, allowing layout designers to complete manual routing operations error-free and DRC-correct every time. Alternatively, the router can cut layout time by automating the entire layout procedure. And, finally, as with Pulsic’s UniRoute™, which offers all these capabilities, it should be production-proven at the most advanced nanometer process nodes.

The industry has learned from the 65nm and 45nm experience that process rules for advanced nodes have reached a degree of complexity that is impossible to handle within existing resource constraints through manual design. The requirements of emerging custom designs will increase the need for simultaneous and automated routing of analog and digital nets within the same design, using the same technology, features, and database for the entire design environment. Eliminating the need for, and cost of, separate tools while also enabling custom design automation and high-capability routing will ensure that custom designers can meet the needs of advanced process nodes.

The semiconductor industry stands on the shoulders of its EDA tool providers and we must deliver the innovations and productivity necessary to feed and nurture the designers that have come to rely on us. It is only through this combination of innovation and productivity that we can provide the sustained value that will serve as the growth catalyst our broader ecosystem thrives on.

Innovation without context is irrelevant, however, so it’s essential that we deliver technology and capability in a manner that can be applied and exploited by the intended user. Through our interactions and discussions with designers, we consistently hear that many EDA design tools have exceeded the core requirements for a majority of the user base. In fact, just last week one of our customers referred to their usage of a “big three” vendor tool as “firing up the space shuttle to go to the corner store for milk.” This excess is understandable, as the market leaders are driven by the most extreme requirements for their (niche) user base working in the smallest geometries with unique needs. What’s tragic is that these cumbersome, overburdened tool flows have become the acceptable paradigm for the entire industry. The result is an ever-increasing gap between the requirements of most users and the features and functionality provided by the market-leading tools.

We believe 2010 is ripe for a new paradigm – one where “less is more.” An approach to tool design that delivers just the right mix of top-notch features and functionality that is squarely aligned with requirements. This concept of elegant, efficient design is what John Tanner embraced when he founded Tanner EDA twenty-two years ago and it’s an approach that we believe is not only relevant — but imperative — today.

Delivering on “less is more” is difficult. Anyone who has tried to distill a presentation to one slide or simplify a complex design knows that it requires more than just skill. One must achieve a deep level of understanding in order to get to the essence of the topic or issue. For EDA products, we think that functionality will not include superfluous features but instead will deliver excellent, tested and well supported solutions. This cannot be achieved in a vacuum; it requires the leveraging of users, partners, and even competitors. We believe this leverage – achieved through models such as “open innovation” (originally coined by Professor Henry Chesbrough) — is essential to achieving and sustaining “less is more.”

The open innovation business model offers a compelling framework for consideration in the EDA industry. With a core principle that ideas and intellectual capital can come from outside the traditional boundaries and connections, open innovation can bring new capabilities and technologies efficiently and effectively. Companies in the broader EDA ecosystem (such as Qualcomm) have already embraced open innovation as a means of effectively bolstering their innovation capacity and effectiveness.

While perhaps not considered a traditional example of open innovation, process design kits (PDKs) offer a congruent model for connecting technologies and intellectual property (IP) from one domain (IC fabrication) to another (design). One perspective on PDKs is that they are simply rule-sets that provide all users with a consistent base of information; effectively eroding opportunity for differentiation within a design. However, further consideration reveals that there are several other dimensions to PDKs where unique IP can be inserted for sustained differentiation.

One such dimension is PDK selection: simply identifying and applying technologies within and across foundries. A working example of this is Tanner’s recent collaboration with Sound Design Technologies (SDT). SDT and Tanner are launching a PDK to allow users to include advanced integrated passives technologies and 3D chip packaging capabilities in their designs. This offers the potential for substantial space savings as well as production and operating cost reductions. The other dimension here is access – where a tool vendor’s use of standard programming languages and opening of PDKs can provide a designer with the access and opportunity to customize and create differentiation.

Productivity is not exclusive from innovation; in fact, we believe that in this new era of doing more with less, designers will require more productivity from their EDA tools if they are to achieve the breakthroughs demanded by their customers. We believe that 2010 will see productivity requirements expand beyond the basics of performance, security and capability. Significant advances in the area of design environments and analog automation will achieve prominence. And designers will be able to use more efficient, focused tools to deliver profound breakthroughs for business and society.

By leveraging the strengths of our award-winning technologies, Apache expands its power and noise product offerings from SoC, analog/mixed-signal, and package/PCB designs to the RTL where greater opportunity for power optimizations can be realized. It’s a real “win-win” for everyone: our Design For Power customers have a more synergistic and complete low-power flow; our employees have expanded opportunities for collaboration and innovation; and the EDA ecosystem maintains a healthy, competitive balance for low-power design.

Low power is an area that affects every aspect of our lives. You can readily see how tiny power numbers can have major impacts in the latest Energy Star guidelines for various appliances and systems. Their chart highlights the typical energy profile and standby power for several common appliances. The line graph denotes the potential energy savings from following the Energy Star specifications; this graph is significant because it shows a system as ubiquitous as the external power supply has more potential power savings than a data center server!

Our customers can easily, and automatically, achieve power reductions in advanced SoC designs of 20 percent or more today, and we expect Apache to continue to lead in these efforts, while providing expanded opportunities for low-power design success. As our CEO Andrew Yang puts it: “This acquisition reinforces Apache’s business and product strategy for a complete offering of advanced power and noise integrity solutions for chip-package-system convergence.”

I have been championing the concept of “aggregation of smaller innovative private companies” to create a bigger footprint and solve bigger problems facing IP and SoC designers below 65 nm for some time. This has many positive aspects, particularly in today’s less-than-ideal economic conditions.

The private company executive still goes up against the big boys, and that requires innovation and determination. By acquiring the technology and skill sets to create a more competitive enterprise, while remaining small and nimble enough to take advantage of opportunities as they arise, the savvy entrepreneur can stay one step ahead.
EDA is also in a true Darwinian situation today – only the fittest companies will survive. Private companies are no longer heading for a quick IPO, or buyout from one of the Big Three (who have problems of their own), so the best option is to be in the marketplace with the best ideas, best service, and best people.

I have observed a continuing trend over the last 10 years of customers not wanting to speak about the technologies they are adopting and forbidding their employees to speak out. It started with the famous “well know microprocessor company” whom you could not even acknowledge having as a customer in your investor presentation. But the trend has extended and is now pervasive. I know for example that Synfora has a customer that went from using the product on one design to using it on 40 designs in one year but they won’t talk. Another one sent an email to the company entitled “PICO results are SWEET” (caps in the original email) and showed that it was now beating hand design in every category but for both of them it would take weeks of begging and hassling for authorization. In another domain, one semiconductor company won a major socket with “a major networking company” but the world will probably never hear of it until they go public and have to disclose customers above a certain percentage. When Certess was acquired by Springsoft earlier in the year, it had 50 active design sites but with a few notable exceptions most were never made public.

There was a time when people were competing for being recognized as early adopters of technology. Companies were proud to author papers in conferences showing how a new concept was being applied on their product development. Engineers would get plaques for publications and those were recruitment tools to get the best graduates to join their companies. Now when you look at user forums such as Deepchip, most of the engineers have to go “anonymous” and carefully disguise who they are before they can even acknowledge using a new technology.

So dear customers, we love you, we work long hours to develop the new technologies that will continue to propel you forward but we need your help to be able to help you. If you want new technologies to be developed and if you want to have investors plow money on your behalf, then we need your help. Come back to the proud tradition of publicizing engineering innovation and adoption. Write about what you do, blog about what you use, accept to be quoted and interviewed and I am sure this will greatly increase the selection of the best technologies….for your own benefit !

http://www.deepchip.com/wiretap/090709.html

With all the exciting new products and new companies, this is not the year to miss DAC!

Hi John,

This is in response to the post from Gayatri Japa on TechGuri. First, Apache did not launch TechGuri; it is not an Apache site. Apache is a blogger on the site, along with Atoptech, Springsoft and Synfora.

TechGuri is an independent blog site offering high tech companies an avenue to blog and create open discussion without the responsibilities of maintaining, hosting, and administrating a blog site of their own. Anyone who blogs or maintains their own site, knows blogging, content creation, and site maintenance all take time. Many companies do not have the in house ability for a full team of bloggers, or have a team of webmasters to manage a separate blog site or page. TechGuri offers what every large company offers with their own bloggers by aggregating the midsize to small companies. TechGuri allows sponsor companies to focus on their core strengths of technology and R&D while giving them an opportunity to communicate to their audience in this new age of social media.

The bloggers on TechGuri are not paid to blog, they are not editors, or journalist and they do not promote themselves in that way. Each blogger is a technical expert from the company s/he is representing. Each company has an opportunity to link directly from their company website to the TechGuri author page and from TechGuri to the bloggers company site. There is no misleading here. The TechGuri site was recently launched, two weeks ago in early June and we are still working on coordinating all the crosslink’s from site to site. We have had several suggestions on how to improve and update the site, which we are currently working on.

TechGuri also offers all companies, whether a blog sponsor or not, an opportunity to blog on the site. There are posts from guest bloggers; currently posted is a post from Erach Desai, on the Unionization of Engineering. Companies are also offered an opportunity to place ads on the site.

As for Gayatri’s comment on “fake ads” and “fake bloggers”. The ads on the site are actual sponsor company ads that lead directly to the sponsors site. The bloggers are employees of the respective company, and blog just like everyone else who blogs. These are actual EDA engineer’s blogging about their specific topic of expertise.

I encourage everyone to drop in and visit TechGuri. Sign up to follow us on Twitter also. There will be new posts weekly by various companies on various topics. I do hope this clears up any misconception of Techguri and Apache’s involvement.

Michelle Clancy
TechGuri, founder and administrator
www.Techguri.com

The outsourcing debate has been ongoing for several years. Prior to the financial meltdown in 2008, however, anyone who questioned the “wisdom” of outsourcing engineering jobs to more lower-cost regions of the world would have been placed on the waiting-list for Guantanamo Bay. Earlier this year, Ron Hira, professor of public policy at Rochester Institute of Technology, touched off a firestorm by penning an opinion piece in EETimes that was titled: The Obama administration promotes outsourcing. Coming from an academic, who would be stereotyped as an O-bot, the piece has especially generated a plethora of comments and critique (O-bot: one who is programmed to robotically carry out the policies of the messiah; credit for “O-bot” goes to Michael Graham, controversial radio talk-show personality in the greater Boston area). And, IBM with their stealth layoffs and concomitant hiring boost overseas has become the poster-child of the sins of outsourcing, with labor unions clamoring for proverbial blood.

As an electrical engineering graduate, I arrived in Silicon Valley in the mid-1980s when semiconductor companies actually had fabs in the valley running three shifts. Being young and idealistic I witnessed with some bemusement the efforts of various traditional labor unions to band together the fab line workers. The senior and seasoned design engineers, my mentors, were adamant in advising us “young guns” not to fall prey to the promises of alleged mobsters and thugs. Rhetoric aside, the more well-reasoned argument was: these are lower-skilled manufacturing jobs that can be more efficiently (read: more cheaply) done overseas. Okay!

When I went to business school, for my technical lobotomy, I distinctly recall the debate over labor unions in our human resources class. Our professor was a typical labor-loving liberal, and most of the section on labor unions was spent demonizing Ronald Regan (for dissing the air-traffic controller strike) and Robert Crandall (long-standing CEO of American Airlines for similar union-busting sins). One undergraduate business major, who annoyingly registered for most MBA courses, piped up this observation: perhaps unions have out-lived their usefulness in a knowledge-based economy? Pretty big picture thinking for an undergrad runt; though I’m not sure what that comment did for his grade! Undoubtedly, the legacy unions (airlines, auto manufacturers, etc.) have largely contributed to the demise of industrial America over the past 20-30 years.

Fast-forward to the late 1990s: the technology economy was on fire, putting an “e“ in front of your company name or product basically gave one “legitimacy”, and dot-coms with flimsy business models were going IPO. Outsourcing engineering jobs overseas had already begun, but largely in the information technology software and business process arenas. These were allegedly the lower-skilled call center and grunt work jobs that could more efficiently (read: more cheaply) done overseas. Hmmm!

Design engineering for the semiconductor and systems industries was largely immune from the outsourcing trend through the technology bubble. The focus then was on revenue growth while profitability was this “old-fashioned, prude” concept. Then came the nuclear winter of technology: circa 2001-2003. The technology IPO market that had labored through 2001 went through a drought through most of 2004. Venture capitalists fortunately found “religion”, and began focusing on profitability. And, like any herd mentality trend (someone actually wrote a book about The Tipping Point!?!), “do you have an India outsourcing strategy” became a necessary check-mark for entrepreneurs looking for funding along Sand Hill Road. The argument was now articulated along pure economic lines: if we can hire four engineers in India (or Russia, or Armenia, or China) for one engineer in the US (or Silicon Valley), we have to do the fiduciary responsible thing for our shareholders, right? Hmmm again!

I have spent the bulk of the last 12 years as a sell-side analyst for design software and semiconductors on Wall Street. Additionally, my friends know that my political leanings usually tend towards the conservative side (or right of center). So, what’s the point of this diatribe; or, in other words, what’s my beef against open borders, globalization, and financial shenanigans in general? Allow me to elucidate:

At heart, I am an engineer (a polished geek, if you will). Having stayed in touch with my industry peers through the years, I have great empathy for what is going on with experienced engineers as a direct consequence of the outsourcing trend.

The financial analyst and journalism vocations have also been impacted by outsourcing; although less visibly so, since there are sea-changes in the underling financial services and media industries.

As a 40+ year old job seeker during multiple stages over the past four years, I have encountered blatant age discrimination (the politically correct HR-speak is “over experienced”; as opposed to over-qualified). And, while I have no desire for the US to adopt French employment laws (read: socialism), I have also fallen victim to the overt abuse of “at will” employment contracts.

Okay (sob, sob), we feel for you dude, but what the heck does this have to do with the outsourcing trend? Market forces are in play, and one can either adapt, or get out of the way!

Here’s the rub! There will always be some new, emerging geography with a growing population of educated engineers willing to work for relative peanuts. Taken to the limit, we will have outsourced all of the engineering jobs (mundane, creative, and innovative), all of the marketing jobs, all of the financial analysis jobs, all of the journalism jobs, etc. The only professions one cannot outsource (yet) will be key service providers (doctors, teachers, firemen, policemen, etc.), sales people (at least for direct sales), and the entertainment, hospitality, and recreation industries. Oh, and the all important, Wall Street suits, money managers and lawyers. Everything we consume (outside of food) will be designed and manufactured at some low-cost location. Just hunky dory!

I don’t study demographic trends as a hobby, but the population in the US does continue to grow. Almost everyone’s retirement funds were slashed in half in 2008. So, for mere mortals, retirement is a more elusive goal. Unless the outsourcing trend is stemmed, we are going to have a middle class revolution on our hands. The “tea party” protests about a month ago will likely prove to be a mild prelude to louder and more frequent protests against capitalism run amuck.

Free-market capitalism has been running amuck since the repeal of the Glass-Stegall Act under Clinton. And, good ole’ George might have slept through some of his classes on corporate finance and leverage at HBS. Amazingly, I actually like the gist of Obama’s corporate tax re-haul proposal: change economically undesirable behavior by modifying the tax code (so companies that outsource more jobs will not be able to expense those jobs for tax purposes). We’ll have to see if Congress goes along with it, and whether this was more public posturing by candidate Obama (stop campaigning, dude; you won!).

The grass roots answer may lie in an “enlightened” form of unionization by rank-and-file engineers in the US. We know that the legacy unions grew too big and powerful, got corrupted, and abused their missions. We can do better. Engineers are eternally pragmatic, and my guess is that most would be willing to “get real” about compensation expectations and rejuvenate their work ethic if there was an opportunity to continue gainful employment.

What’s your answer to this dilemma?

Yes, an “enlightened” form of unionization makes sense

Change is here and Obama’s corporate tax re-haul plan will work

No one’s listening: let’s just take to the streets

What dilemma! The US cannot fight the forces of globalization and economics; go bag groceries, dude!