Tensilica Says Industry On Road to a ``Sea-of-Processors'' As SOC and Microprocessor Development Worlds Converge.Business Editors/High-Tech Writers SANTA CLARA, Calif.--(BUSINESS WIRE)--Oct. 17, 2001 CEO Chris Rowen To Unveil Industry Vision at Microprocessor Forum 2001 Tensilica(R) Inc. will today describe the inevitable, fundamental shift that is expected to occur in microprocessor and System-on-Chip (SOC) development in the coming decade at "Microprocessor Forum 2001." In a keynote speech entitled "Bridging the Microprocessor -- SOC Design Gap," Tensilica's CEO, Dr. Chris Rowen, will outline the rising need for processors that can be easily customized and optimized for specific applications, as well as seamlessly integrated onto SOC devices. This impetus will drive the industry away from standard, fixed instruction set processors (language) Instruction Set Processor - (ISP) A family of languages for describing the instruction sets of computers. ["Computer Structures: Readings and Examples", D.P. Siewiorek et al, McGraw-Hill 1982]. and traditional sea-of-gates ASIC approaches. Dr. Rowen's "Sea of Processors?" vision will be presented this evening at 5:30 p.m. at the Fairmont Hotel in San Jose, Calif. "Today's standard, fixed instruction set processors cannot be optimized for their target applications unless designers invest immeasurable resources into developing accompanying complex ASIC blocks," explains Dr. Rowen. "This process is not only time-consuming and costly, but is also completely impractical for SOC designs with multiple processors. The industry needs a microprocessor environment that enables customers to combine both standard functions and their own customized logic onto the processor. Moreover, by automating this process, designers will have a smooth path to multiprocessor integration on SOCs." Automation Required to Achieve a "Sea of Processors" Dr. Rowen believes the industry needs a development environment that enables customers to tailor the CPU for their target application, adding their own customized instructions as needed to the processor, and alleviating the need for complex ASIC solutions. Furthermore, this process must be fully automated, allowing designers to generate their desired CPU at virtually the click of a button. The end result will be a microprocessor/SOC design world that maximizes chip real estate and dramatically reduces design time. Dr. Rowen's presentation will outline five stages of design enlightenment that are expected to occur in the microprocessor and SOC design process. These stages include: -- Stage One: System OEMs operating at the first evolutionary stage do not employ ASICs or SOCs due to the high cost of EDA tools and non-recurring engineering (NRE) costs, as well as the lack of sophisticated technical knowledge needed to design SOCs. These companies purchase off-the-shelf integrated circuits and place them on circuit boards. Chip customization is limited to burning EPROMs and PLDs. The focus is on board-level design. -- Stage Two: Designers begin developing primitive SOCs with fixed-architecture cores. System OEMs at this level often create SOCs by replicating the functions of existing boards on a chip. Therefore, they use fixed-function duplicates of processors already used on PC boards. This enables them to work with familiar processor architectures and to preserve legacy code. -- Stage Three: Configurable cores come into play. After learning to build simpler SOCs, system OEMs become ready to start tailoring processors for applications. The simplest configuration parameters are selectable blocks, such as floating-point units, and cache sizing (none, 8k, 16k, 32k). The core processor architecture is not altered, preserving legacy code. -- Stage Four: Designers begin using extensible processors. Once system OEMs become acclimated to the idea of customizing processors on SOCs for specific applications, the next step is to start extending the processor with special instructions that significantly improve the processor's performance. By making improvements to certain parts of the code using these new instructions, designers can achieve as much as 100 times the performance of general-purpose processors that cannot be optimized for every application. Tensilica's current product line targets SOC developers operating at this stage of design enlightenment. -- Stage Five: Extensibility becomes automated. At this stage, the microprocessor fades from view. System developers write high-level code in C or C++ that describes what needs to be done. Development tools analyze this code, define a processor with instructions that can optimally run the code, and then compile the C using a specially tailored compiler that uses the tailor-made processor. Because the tools to support this design approach do not yet exist, no one currently designs SOCs or systems at this level. Assessing the Industry: Where are we now? Dr. Rowen believes that most of the electronics industry continues to operate somewhere between Stage One and Stage Two, and a few companies have started to employ Stage Three design techniques. However, the only way to unlock the full potential of SOC technology is to operate at Stage Five, which gives design teams a practical way of fully realizing the promise of microprocessors, while eliminating the expensive and error-prone ASIC design methods currently in use. In this future world, individual processors will cost millicents and processor operations will consume femtowatts. In such a world, the total annual processor output could reach 100 billion per year, with 99 percent of all processors shipping as integrated functions on SOCs devices. About Chris Rowen, Ph.D. Dr. Chris Rowen is president and chief executive officer of Tensilica. He founded Tensilica in July 1997 to develop and proliferate automatic generation of application-specific microprocessors for communication and consumer systems. He was a pioneer in the development of RISC architecture at Stanford in the early 80s and helped start MIPS Computer Systems Inc. in 1984, where he served as vice president for Microprocessor Development. Most recently, he was vice president and general manager of the Design Reuse Group of Synopsys Incorporated. He received a BA in physics from Harvard University and MS and PhD in electrical engineering from Stanford University. About Tensilica Tensilica was founded in July 1997 to address the fast-growing market for configurable processors and software development tools for high volume, embedded systems. Using the company's proprietary Xtensa Processor Generator, system-on-chip (SOC) designers can develop a processor subsystem hardware design and a complete software development tool environment tailored to their specific requirements in hours. Tensilica's solutions provide a proven, easy-to-use, methodology that enables designers to achieve optimum application performance in minimum design time. The Company has over 140 engineers engaged in research, development, and customer support from its offices in Santa Clara, Calif.; Burlington, Mass.; Princeton, N.J.; Houston, Texas; Oxford, U.K.; Stockholm, Sweden; Taipei, Taiwan, R.O.C.; and Yokohama, Japan. Tensilica is headquartered in Santa Clara, Calif. (95054) at 3255-6 Scott Boulevard, and can be reached at 408/986-8000 or via www.tensilica.com on the World Wide Web. Note to Editors: "Tensilica" is a registered trademark and "Sea of Processors" is a trademark belonging to Tensilica Inc. All other trademarks and registered trademarks are the property of their respective owners. Tensilica's announced licensees are, in alphabetical order, Berkeley Wireless Research Center, Broadcom, Cisco Systems, Conexant Systems, Fujitsu Ltd., Hughes Network Systems, Ikanos Communications, JNI. |
|
||||||||||||||||

Printer friendly
Cite/link
Email
Feedback
Reader Opinion