2011: The Year of U.S. High-Level Synthesis Deployment

High-level synthesis has been in production use in Asia for nearly a decade. And, while you’ve been able to buy products from leading Japanese and Korean vendors at your local Best Buy for several years, U.S. companies haven’t widely deployed the technology … yet. That is all about to change.

It’s no longer about if high-level synthesis is viable –– it’s about how to deploy high-level synthesis. Companies are putting together laundry lists of requirements to get their design flow moved to a higher level of abstraction. This means that their design requirements need to be handled by high-level design tools. Sure, quality of results is always important, but it is also about tool maturity, tapeouts and practical solutions to real-world design problems at a high level of abstraction.

Why should companies move to a higher level? They are telling us the reason every day: To improve productivity. U.S. productivity numbers have been steadily increasing (1) as the market continues its slow recovery. Staying competitive in the global market without increasing costs is critical to keeping the competitive edge. By increasing productivity, design companies can get more done at a lower cost and drive growth and profitability numbers. High-level synthesis delivers a game-changing 10x productivity increase for new design projects and even more for IP reuse.

One key enabler of the change at the methodology level is the full adoption of SystemC. The fact that even the companies that believed SystemC would “die in a few years” have changed their tune and are trying to rollout their own SystemC tools.

SystemC has been confirmed as the standard for high-level design around the world and now in the U.S. More U.S. design companies are on board, too, as evident by the OSCI membership list (2) that now includes U.S. companies such Broadcom, Intel, Texas Instruments and Qualcomm.

Sure, there are still a few companies trying to push “ANSI-C” or their own flavor of a language but they will be forced to move or simply pass on.

Tool companies that claim SystemC support will have to prove their support is real, not just support for the datatypes. These SystemC models are being pushed in the largest U.S. companies for virtual system prototyping, high-speed verification, high-level synthesis and more. To utilize SystemC models, tools will need to support the full language and nothing short of it. Many companies are scrambling to do this right now though they’ve made their announcements nearly a year ago.

The drive for growth is a powerful one and the stars have aligned with key technologies in place for the U.S. market to take full advantage of the benefits of ESL and high-level synthesis.

 

 

http://www.gabeoneda.com/openchannel/2011-year-us-high-level-synthesis-deployment