Ubicom Unveils Groundbreaking IP3023 Wireless Network Processor Based on Revolutionary ``MASI'' Architecture.Business Editors/High-Tech Writers MOUNTAIN VIEW, Calif.--(BUSINESS WIRE)--April 15, 2003 Entirely New Architecture Category Results from Nearly Five Years of Work in Software I/O (Input/Output) The transfer of data between the CPU and a peripheral device. Every transfer is an output from one device and an input to another. See PC input/output. I/O - Input/Output Ubicom(TM), Inc., a leading provider of wireless network processor and software platforms, today broke new ground in the wireless network processor market with the announcement of its landmark IP3023(TM) processor. The culmination of nearly five years of work in software I/O, the IP3023 is the first in the IP3000(TM) family of wireless network processors, and has led to the definition of an entirely new microprocessor architecture category, called the Multithreaded multithreaded - multithreading Architecture for Software I/O (MASI MASI Mach Airspeed Indicator MASI Manhattan Sites (US National Park Service) MASI Multithreaded Architecture for Software I/O MASI Member of the Architectural and Surveying Institute ). Analogous to the way CISC (Complex Instruction Set Computer) Pronounced "sisk." The traditional architecture of a computer which uses microcode to execute very comprehensive instructions. and RISC RISC in full Reduced Instruction Set Computing Computer architecture that uses a limited number of instructions. RISC became popular in microprocessors in the 1980s. transformed desktop and high-performance computing High-speed computing, which typically refers to supercomputers used in scientific research. , MASI is expected to have a similar impact on the wireless networking See wireless network. market. "We tasked some of the best minds in the engineering world with developing a new microprocessor designed from the ground up for software I/O and wireless networking," said Bulent Celebi, president and CEO (1) (Chief Executive Officer) The highest individual in command of an organization. Typically the president of the company, the CEO reports to the Chairman of the Board. , Ubicom, Inc. "This required an entirely new architectural approach, unlike anything that could be achieved using traditional CISC or RISC techniques. While MASI draws upon many age-old engineering principles, it is the combination of these principles that has enabled us to deliver the industry's first fully optimized wireless network processor." MASI: A Revolutionary Architecture That Will Transform Wireless Networking The MASI architecture is designed specifically to enable high-performance execution of software I/O. Software I/O implements all communications and control functions in software, eliminating costly dedicated hardware I/O that can account for as much as 75 percent of a chip's die area. In addition, because software is fundamentally easier to change than hardware, software I/O can be easily adapted and upgraded as applications demand or as standards evolve. "Wireless networking requires an architecture that is optimized from the ground-up for packet processing -- not one that is encumbered Encumbered A property owned by one party on which a second party reserves the right to make a valid claim, e.g., a bank's holding of a home mortgage encumbers property. by the features and functions needed for general-purpose computing," said Sanjay Iyer, senior analyst, The Linley Group. "Ubicom's MASI architecture, as implemented in the IP3023 wireless communications wireless communications System using radio-frequency, infrared, microwave, or other types of electromagnetic or acoustic waves in place of wires, cables, or fibre optics to transmit signals or data. processor, employs intelligent simplifications and a radical hardware-software partitioning to deliver a high-performance wireless solution that reaches low cost and power levels unattainable by conventional general-purpose communications processors." A variety of well-known engineering principles must come together to achieve high-performance software I/O. It is the combination of these techniques that makes MASI truly revolutionary. The architecture is characterized by: -- Determinism: MASI features deterministic execution, allowing developers to exactly time a software routine. -- Multithreading Multitasking within a single program. It allows multiple streams of execution to take place concurrently within the same program, each stream processing a different transaction or message. : MASI is a multithreaded architecture that, when combined with its deterministic nature, allows each thread to operate independently with no impact from conditions among threads. As a result, each thread can be seen as a separate processor. In addition, each thread is programmable, allowing systems to be easily upgraded in the field as standards change. -- Memory-to-memory: MASI requires a memory-to-memory architecture that enables the processor to move "inputs" directly to memory and "outputs" directly from memory to the I/O pins in real time. -- Bit Manipulation Processing individual bits within a byte. Bit-level manipulation is very low-level programming, often done in graphics and systems programming. : MASI requires strong bit manipulation instructions to allow efficient access of individual I/O pins. IP3023: The Ideal MASI Implementation for High-Performance Wireless Applications The IP3023 is Ubicom's latest implementation of the MASI architecture, and features a mature second-generation instruction set to make it the performance leader in Ubicom's growing product portfolio. The device builds upon the successes gained with its first wireless network processor, the IP2022(TM), achieving a 10x performance improvement, as well as a number of industry firsts. The IP3023 features: -- Eight-Way Multithreading: The IP3023 chip features eight-way deterministic, instruction-level hardware multithreading enabling it to handle eight different execution threads in its pipeline simultaneously. This allows it to act as eight virtual microprocessors, with as many as six threads dedicated to software I/O functions. In addition, a 64-entry scheduling table enables the developer to assign a different, flexible number of MIPS (Million Instructions Per Second) The execution speed of a computer. For example, .5 MIPS is 500,000 instructions per second; 100 MIPS is a hundred million instructions per second. to each hard real time (HRT HRT abbr. hormone replacement therapy Hormone replacement therapy (HRT) Also called estrogen replacement therapy, this controversial treatment is used to relieve the discomforts of menopause. ) thread. Each thread can have a speed ranging from 0 - 250 MHz (MegaHertZ) One million cycles per second. It is used to measure the transmission speed of electronic devices, including channels, buses and the computer's internal clock. A one-megahertz clock (1 MHz) means some number of bits (16, 32, 64, etc. , assigned in 3.9 MHz increments. -- Zero-Cycle Context Switching Same as task switching. : A zero-overhead scheduler enables multithreading that is penalty-free. Not only do developers get to take advantage of up to eight virtual processors, but they can also avoid the typical switching penalties of general-purpose processors. -- Memory-to-Memory: The IP3023 features a memory-to-memory architecture that processes data directly in memory using a single instruction that reads, modifies, and writes the data in a single cycle. Because packet-processing operations need only access data once, the use of a single instruction is significantly more efficient, saving valuable clock cycles and reducing overall code size. This eliminates the load/store instructions required by desktop computing architectures, and found in general-purpose operating systems Operating systems can be categorized by technology, ownership, licensing, working state, usage, and by many other characteristics. In practice, many of these groupings may overlap. . As a result, one instruction in Ubicom's new processor does the work of five instructions on a comparable platform. -- Strong Bit Manipulation: The IP3023 features strong bit manipulation that allows direct access of pins from the CPU CPU in full central processing unit Principal component of a digital computer, composed of a control unit, an instruction-decoding unit, and an arithmetic-logic unit. core, and the efficient manipulation of packet headers. -- Optimized Instruction Set: The IP3023 features an instruction set designed specifically for wireless networking. As a result, it uses just 41 instructions, eliminating much of the overhead associated with traditional approaches and reduces memory requirements by up to 95 percent. This also eliminates the use of costly on-chip caches lending to very low latency. With on-chip memory, the device can achieve 1 gigabyte/second of memory bandwidth. -- Deterministic Operating System: The IP3023 is optimized to run Ubicom's, deterministic operating system, which is structured so that interrupts can be managed without ever turning them off. IP3023: Building on the Successes Achieved with the IP2000 Family While the IP2000(TM) family processors remain ideal for a variety of wireless networking applications, the IP3000 family delivers the added horsepower needed to implement higher data rate protocols such as 802.11a/g, and to drive high-performance wireless applications like gateways, routers and multiprotocol access points. Ubicom has increased the number of threads supported from two on the IP2022, to eight on the IP3023, allowing the processor to support more software I/O simultaneously. Multithreading also enables zero-cycle context switching, an improvement over the already revolutionary 3-cycle context switching supported on the IP2022. The following table summarizes the key features of the IP3023:
Feature IP3023
-------------------------- --------------------------
Threads 8
Context Switch 0 clocks
Data Path 32-bit
Clock Speed 250 MHz
Code RAM 256 kilobytes
Data RAM 64 kilobytes
Process Technology 0.13-micron
MII Ports 4
SerDes Units 2
External Memory Controller SDRAM
Packaging, Pricing and Availability The IP3023 will be housed in a 208-pin plastic quad flat package See PQFP. (PQFP (Plastic Quad Flat Package) Refers to many varieties of QFP chip packages, which are molded in plastic. See QFP. ). Ubicom expects to sample the device in Q22003, with production slated for 2H2003. The IP3023 will sell for $12 in 100,000-unit quantities per month. Please visit Ubicom's Web site at www.ubicom.com for more information. About Ubicom, Inc. Ubicom, Inc. develops and markets wireless network processor and software platforms that enable all electronic devices to be connected to each other -- securely, cost-effectively and transparently. With headquarters in Mountain View, California For the census-designated place, see Mountain View, Contra Costa County, California. For other places called "Mountain View", see . Mountain View is a city in Santa Clara County, in the U.S. state of California. The city gets its name from the views of the Santa Cruz Mountains. , Ubicom also has offices in Southern California, as well as Belgium, Taiwan and Hong Kong. For more information, visit www.ubicom.com. Editors' Notes: -- A photo, quote sheet and analyst reference sheet on the IP3023 are available. Please call Joany Draeger, Tanis Communications, 650/365-3395 or e-mail joany@taniscomm.com. Ubicom, IP2000, IP2022, IP3000 and IP3023 are trademarks of Ubicom, Inc. All other trademarks are the property of their respective holders. |
|
||||||||||||

Printer friendly
Cite/link
Email
Feedback
Reader Opinion