Archive for 2007

I must say I’m very much enjoying blogging about EDA standards. Not only is the topic interesting and important to my customers, but equally so to my competitors. In the short time that I’ve been blogging, I’ve received insightful comments, intriguing questions, and strongly-stated positions.

One thing I’ve really gotten a kick out of is some of the ratings. A creature named webwasher.Cadence.com has visited my blog and left behind but a single star. I think maybe it’s a compliment.

Last weekend, I had the pleasure of attending the 2007 IEEE Standards Association awards ceremony in Hallandale Beach, Florida. Among the many deserving recipients of this year’s awards, there are three gentlemen that I would like to personally congratulate. They are colleagues of mine in the standards world and have made a positive impact on the IEEE Design Automation Standards Committee (DASC).

The Ron Waxman DASC Meritorious Service Award is named after Ron Waxman, a founder of the DASC, in recognition of his many years of leadership and service to IEEE and international standards. Ron told me that he could not have foreseen all the wondrous events in the EDA industry of the past two decades, but he did have a vision. Those who followed him all worked to make DASC the significant body it has become, advancing design automation standards which in turn benefit design practices. Ron can be proud of how the DASC works in harmony with many ad hoc industry standards organizations so the resulting IEEE standards can satisfy the needs of all, both politically and technically.

Gabe Moretti is the first recipient of the annual Ron Waxman DASC Meritorious Service Award. This award recognizes commendable accomplishments by a DASC member who puts forth extra effort to further the DASC and its mission. Gabe has delivered a high level of service to the DASC over many years and contributed significantly to several key activities including VHDL 1076 and Verilog 1364. He took the DASC beyond its original VHDL-only focus, giving it a much wider and compelling scope for design automation standards. Gabe actively promoted IEEE corporate standards, helping bring the DASC to a new level and greatly rejuvenating it.

Victor Berman, on behalf of the DASC, accepted the IEEE Standards Association Sponsor Award for “Outstanding Contribution to Corporate Standards Development.” This award is presented to IEEE standards sponsors that have shown exceptional support of the entity standards development process. To date, DASC has sponsored half of the current entity-based standards and almost one-third of the active entity projects. It forged bonds with three other standards organizations to create a proven path for their standards to become IEEE entity standards. DASC sponsorship of these standards not only helped pioneer the entity standards process, but also promoted it throughout the design automation industry as a highly effective and efficient means for producing market-relevant standards.

Please join me in congratulating Ron Waxman, Gabe Moretti, and Victor Berman.

EDA standards blog
DASC IEEE design automation standards Ron Waxman Gabe Moretti Victor Berman IEEE awards

Elections are underway for Chair and Vice Chair of the Design Automation Standards Committee (DASC). Nominations are due this Friday, November 30, 2007.

Any member of the DASC can nominate individuals for these important positions. If you are not a current member of DASC, you can join anytime. Visit the DASC home page for a membership application. (Or, contact me – I’m the DASC Treasurer.)

Please send nominations directly to the DASC elections officer, Dr. Peter J. Ashenden, at peter @ ashenden.com.au . As he receives nominations, he will confirm with the nominees that they are eligible and willing to stand.

Nominating qualified individuals for DASC officers is one way you can help keep DASC a vibrant standards-setting organization.
EDA standards blog
DASC IEEE design automation standards

The big standards news in July was about the Virtual Socket Interface Alliance (VSIA) closing operations and moving their successful products to other industry organizations. I’ll bet VSIA did a lot of soul searching before making this decision, which I think was the right one. I give them much credit for having the courage to take what must have been a difficult step.

This could have been the end of the line for a standards-based solution for protecting IP. (In this context, IP refers to the intellectual property “building blocks” used in designing a complex chip.)

However, VSIA decided to transfer their work on an IP Encryption standard to the IEEE Standards Association. They selected the IEEE’s Design Automation Standards Committee (DASC) to become the new home for their budding standard.

Rights to the work were assigned to the IEEE, and a Study Group was formed with DASC as its sponsor. Dedicated participants in the Study Group, under the leadership of a very capable chair, continued the work almost as if nothing had happened.

Today, the IEEE DASC approved a new Project Authorization Request (PAR) for the IP Encryption standardization project. Dubbed P1735, the project has now been officially sanctioned by its sponsor. This is an important, early step in the process of creating an IEEE standard.

What happens next? The PAR will be submitted to the IEEE New Standards Committee (NESCOM) for approval. In the meantime, the Study Group can continue making progress, with a bit more wind behind their sails.
EDA standards blog

IEEE IP standards IP encryption VSIA DASC

One of my colleagues asked me, “Karen, how does one go about creating a new standard?”

There are different ways to start a standard, depending on your goals. Regardless of the path you take, the first step is to find other interested parties (both suppliers and customers) who are willing to put forth the effort to produce and adopt a standard.

Open source standards are the fastest to create. They are managed by a single entity and everyone is allowed to suggest changes and enhancements. The managing entity decides what changes go into subsequent versions of the standard. Linux, OpenAccess, and Si2′s Liberty TAB (for Synopsys’ Liberty standard) are good examples. A web-based system is needed, such as that provided by SourceForge or Collabnet or one developed by the managing entity.

Formal standards can be slower, but they come with official accreditation, which can lend more credibility to them. Organizations like the IEEE Standards Association, ISO, IEC and Accellera have well-established policies and procedures for producing standards. These documents are readily available, and members of the organizations are generally eager to help get a new standard started.

Usually, the overall process begins with the group of interested parties getting together and defining a strawman of what the standard would be like – mostly a requirements or objectives document. Then, they decide on open source or a formal standards organization. Technology donations are made (if there are any to be had), and the group produces the completed standard, following the open source or formal procedure. Depending on the complexity of the standard, it can be ready for market adoption in as little as 6 months or several years.

One of the most important goals of producing a standard is to ensure that the standard comes out in a timely fashion. In our fast-paced industry, it’s a sad situation when technology moves faster than a standard in development, rendering it obsolete before it is realized.

EDA standards blog
formal standard open source IEEE ISO IEC Accellera Liberty OpenAccess

There are two main types of committees that create standards: individual and entity. Which is better for producing standards today?

In an individual-based committee, the group is comprised of people representing themselves or the interests of their employers. Each person gets one vote. Membership fees are low or nonexistent. Services such as technical editing and project management are provided by volunteers or paid for by benevolent contributors.

By contrast, an entity-based standards committee is made up of people representing entities such as companies and organizations. Many individuals from a single entity can work on the standard, but each entity is allowed only one vote, cast by its designated representative. The entities are assessed fees to fund services that are required by the group.

I believe that entity-based standards committees make the most sense today because they address the immediate needs of the market. They are:

- staffed by participants committed to a positive outcome

- timely, addressing contemporary standards requirements

- available sooner, often in a year or less

- well-funded (an important ingredient to any successful project)

- less likely to suffer from dominance by a single company

- not averse to discussing business and political issues

- able to provide a level playing field for standards work

I predict that for the foreseeable future, the most widely adopted and useful standards will come from entity-based committees.

EDA standards blog
IEEE entity standard individual standard

Out of the clear blue sky, I recently heard from a colleague with whom I had worked 20+ years ago. He had stumbled across my face while searching for some silicon technology information, and he decided to send me an email to say hello. It struck me that I had received an “un-obituary” – a notice about an old friend who is alive and well. Un-obituaries are a treat and bring a warm feeling of being connected to the human fabric. Like the “un-birthdays” of Alice in Wonderland, I do enjoy celebrating “un-obituaries”.

EDA standards blog

In case I didn’t see you at the 20th EDA Developers’ Interoperability Forum, here are some of the happenings:

- an impressive, live demo of 5 companies interoperating around the OpenAccess-based IPL (Interoperable Pcell Library) for analog/full-custom flows

- details of how the VMM (Verification Methodology Manual) harnesses the power of SystemVerilog, as proven by 100s of teams who’ve used it globally and an ecosystem of tool, IP, training, and service providers

- a keynote speech from Michael Keating, co-author of the “Low Power Methodology Manual”

- the first public presentation of the new SOI (Silicon-on-Insulator) Consortium

- updates, technical usage information, and 8 EDA and IP vendors’ support testimonials for the UPF (Unified Power Format) standard which has formed the basis for the IEEE 1801 standard project

- new features in the Liberty standard for low power and variation-aware modeling

- prizes, networking, and camaraderie

The event was interesting, informative and fun (despite my dream the night before that I moderated a session in my pajamas).

I hope to see you at the 21st Interoperability Forum in April, 2008!

Presentations from the 20th Interoperability Forum are now available.

EDA standards blog

About a year and a half ago, the electronic design automation industry, along with its customers designing low power integrated circuits, began efforts to create a standard to describe low power design intent for implementation, analysis, and verification.

Today, there are two standards in play: UPF (Unified Power Format) and CPF (Common Power Format). UPF was developed by Accellera using donations from seven companies, while CPF was donated to Si2′s Low Power Coalition by Cadence Design Systems.

Just about everyone I know wants there to be a single standard. Supporting two standards is costly to suppliers and customers alike.

Is there any hope that there will be a single standard for low power design? I, the optimist, say “Yes”!

Within the IEEE, a working group is actively moving towards ratification of IEEE standard 1801. Participants in the group are from both Accellera’s UPF technical subcommittee and Si2′s Low Power Coalition. The results of the Low Power Coalition’s CPF/UPF comparison study have been shared with the IEEE working group to help ensure completeness of their standard-in-progress, with UPF forming the basis of their work.

I expect that when the 1801 standard is sanctioned by the IEEE, the industry need for a single standard will be fulfilled. The sigh of relief should be audible.

(For more information about how the UPF standard was created, see the August/September issue of Chip Design Magazine: “The Unified Power Format – A Standard Done the Right Way”.)

EDA standards blog
IEEE 1801 UPF CPF low power standard

Next week, Thursday, October 25, the 20th electronic design automation (EDA) Developers’ Interoperability Forum will be held in Silicon Valley. The keynote speaker will be Michael Keating, Synopsys Fellow and lead author of the recently released “Low Power Methodology Manual” (LPMM).

The Forum will include a live demonstration of tool interoperability by members of the Interoperable Pcells Library (IPL) initiative, a session on the Unified Power Format (UPF), and the first public presentation of SOI Industry Consortium.

The event is open for everyone to attend at no charge. It will be held at the Sun Conference Center at Agnews Historic Park in Santa Clara CA from 9:00am to 4:30pm. For more information, directions, and to register, visit:
20th Interoperability Forum Registration

EDA standards blog
UPF IPL