News and New Products

Accellera, Spirit Consortium merger hints at future of ESL design

The merger aims to bridge the language-based design and IP-assembly worlds and will exploit the fact that the two organizations have been working in complementary areas of front-end design and verification.

By Ron Wilson, Executive Editor -- EDN, 6/11/2009

Accellera, the standards-developing organization responsible for driving the creation of such IEEE standards as 1364 Verilog, 1076-2008 VHDL, 1497 Standard Delay Format, 1603 Advanced Library Format, and 1801 Low-Power Design Standards, as well as the SCI-MI simulation modeling interface, and the SPIRIT Consortium, primarily known for the IP-XACT IP management and integration system, today announced that the two groups will merge. Details of the new organization will be announced at or near the Design Automation Conference (DAC) next month.

Shrenik Mehta, chairman of Accellera, said the merger would exploit the fact that the two organizations have been working in complementary areas of front-end design and verification, but with a common goal. "We both seek to achieve IEEE ratification for the standards we propose," Mehta said. "We can learn from each other's experience in working with the IEEE, and share each other's resources in this area."

The merger will produce both organizational efficiencies and a single point of contact between the various standards efforts and the new, streamlined and rationalized IEEE standards process. But Mehta and SPIRIT Consortium President Ralph von Vignau both emphasized that the importance of the merger went far beyond efficiency.

Von Vignau explained that in the world of IC front-end design, there have been two dominant paradigms: one based on using formal languages such as Verilog, System Verilog, or SystemC to describe the design and then employing synthesis to produce the chip, and the other based on qualifying and integrating previously developed IP (intellectual property) blocks to create the chip.

"The time has come to bridge the language-based design and IP-assembly worlds," von Vignau said. "As the necessity for first-time success in IC design has increased, the distinction between these two views of how to achieve it has begun to dissolve. Eventually, success requires a single, global view of the design process that incorporates both language-based expression of the design and orderly reuse of IP."

"Our vision is that these two views will become facets of the same development methodology," Mehta agreed.

Integration of the system-level language approach to design with the IP-XACT reuse process could significantly benefit supporters of both ideas. By integrating IP-XACT's XML-based data structures and tools into the same framework with system-level design and verification languages—for instance, by creating a library system driven by IP-XACT—the new organization could finally produce a useful way for front-end designers to explicitly include reused IP in their system-level views of the design. Additionally, IP-XACT could be invaluable in generating test benches for the full design. On the other side, by using language-based tools to help generate IP-XACT structures, IP developers could finally move toward a tool-set equal to the task of explicitly defining IP blocks.

The merger could also help greatly with the rather slow adoption of IP-XACT, von Vignau suggested. "The tools to support the SPIRIT methodology have been slowly emerging. But working with Accellera could make IP-XACT much more deployable than it has been to date."

The board of the new organization will include the board members from both Accellera and SPIRIT. Mehta said that the initial goal for the new board will be structural: to combine and reorganize the two groups' technical committees. After that, the board will turn its attention to what the organization's goals for the future should be. As was true of both Accellera and SPIRIT, the new group welcomes input and participation from all interested parties.