Printer Friendly
The Free Library
14,506,237 articles and books
Member login
User name  
Password 
 
Join us Forgot password?

[0] PMC-Sierra Unveils Industry's Highest Performance MIPS Multiprocessor Architecture.


Business Editors, Technology Writers

NOTE TO MEDIA: Graphic is available in a Smart News Release(TM) on

Business Wire's Home Page at www.businesswire.com and at

www.newstream.com

SAN JOSE San Jose, city, United States
San Jose (sănəzā`, săn hōzā`), city (1990 pop. 782,248), seat of Santa Clara co., W central Calif.; founded 1777, inc. 1850.
, Calif.--(BUSINESS WIRE)--June 11, 2001

RM9000x2(TM) Integrated Multiprocessor Features Dual GigaHertz CPUs to

Address the Performance and Low Power Requirements of the

Communications Market

Today at the Embedded Processor A CPU chip used in a system other than a general purpose workstation, desktop or laptop computer. Such chips are used by the billions every year in a myriad of products. See embedded system.  Forum, PMC-Sierra (Nasdaq:PMCS PMCS PMC Sierra (stock symbol)
PMCS Project Management Control System
PMCS partial mission-capable, supply (US DoD)
PMCS Preventive Maintenance Checks & Services
PMCS Professional Military Comptroller School
) introduced the architecture for the RM9000x2(TM) integrated multiprocessor, the company's next generation high performance 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.  processor designed for the communications market.

The RM9000x2 is a scalable multiprocessing architecture that solves the industry-wide problem of slow data transfers between processors in cache coherent systems and offers greater throughput efficiencies over single 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.
 products. Anchored by dual CPU cores running at 1 GHz, the RM9000x2 achieves maximum performance while drawing only five watts of power. Using 0.13 micron high performance process, the GigaHertz CPU speed and low power is achieved. The twin CPUs connect to high speed memory and I/O interfaces through a multiport, shared memory fabric. High speed 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
 connectivity, which includes a very high performance 500 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.  HyperTransport(TM) bus interface, positions the RM9000x2 to target key applications such as core routers, edge routers, remote access products and enterprise servers.

"Networking is inherently a parallel operation and PMC-Sierra has adapted its microprocessor design to the requirements of high speed processing within networking equipment," stated Tom Riordan, vice president and general manager of the MIPS Processor Division at PMC-Sierra. "The RM9000x2 architecture is the foundation for future products that will provide greater processing power through additional CPU cores and higher performance I/O interconnect designed for networking equipment."

"PMC-Sierra's MIPS design team has a long standing history of delivering leading-edge processor architectures," said Markus Levy, senior analyst at MicroDesign Resources. "At 1 GHz, the RM9000x2 dual processor establishes a new standard for high performance networking, eliminating the single CPU architecture with a unique and highly integrated system solution poised to solve the multiprocessing bottleneck."

RM9000 CPU Subsystem Is Optimized For Performance

At the heart of the CPU subsystem are two RM9000 64-bit MIPS CPU cores that are compatible with the MIPS(TM)-64 instruction set architecture. Each core has a sophisticated cache architecture which includes high performance level 1 instruction and data caches, as well as 256 Kbit of level 2 cache See L2 cache.

level 2 cache - secondary cache
, providing a total of 512 Kbit of level 2 cache for both cores. The large level 2 cache is ideal for storing networking related data structures such as routing look-up tables. The level 1 cache See L1 cache.

level 1 cache - primary cache
 accesses are at the core frequency and with the RM9000x2, access to the tightly coupled level 2 cache is only five CPU cycles. The ability to access the cache so quickly allows the CPUs very fast access to instructions and data and gives the RM9000x2 best-in-class performance for a GHz multiprocessor. Fast Packet Cache(TM) mode allows bypassing of level 2 cache on a per page basis which greatly improves packet processing performance.

The CPU cores have the flexibility to function either with or without hardware cache coherency and allow the RM9000x2 to address a variety of applications. In a non-coherent configuration, the CPUs can be separated into control and data plane processors that provide the foundation for an edge router. A cache coherent CPU subsystem extends the processing power of a single CPU and delivers the high performance needed by a core router exception processor or an enterprise server.

The dual RM9000s are connected to each other by a CPU switch and enable cache transfers between the processors at the CPU core frequency. This architecture solves an industry-wide problem of slow data transfers between processors in cache coherent systems by delivering 51.2 Gbit/s of inter-CPU bandwidth. To accelerate the multiprocessing capabilities, a five-state cache coherency protocol is used. The five state Modified/Owned (Modified-Shared)/Exclusive/ Shared/Invalid (MOESI MOESI Modified Owner Exclusive Shared Invalid ) protocol extends the functionality of the standard Modified/Exclusive/ Shared/Invalid (MESI MESI Modified Exclusive Shared Invalid (states of cache memory)
MESI Modified Exclusive Shared and Invalid
) protocol to permit one processor to share modified data from the other processor's cache. A packet processing application could benefit from this by using one CPU to do the initial processing and the second CPU to finish processing while keeping the packet within cache. Sharing modified cache lines minimize access to the main memory, further increasing the RM9000x2 processor performance. Full hardware I/O coherency co·her·en·cy  
n. pl. co·her·en·cies
Coherence.

Noun 1. coherency - the state of cohering or sticking together
coherence, cohesion, cohesiveness
 is supported over the HyperTransport and SysAD interfaces and enables I/O devices access to coherent memory without software intervention.

Shared Memory Fabric Switches the Data

The shared memory fabric is a high bandwidth data switch that improves overall performance by supporting concurrent data transfers on multiple ports. The shared memory fabric, for example, will allow a CPU access over SysAD at the same time that a HyperTransport peripheral is accessing the main memory. Switching data concurrently eliminates the costly bottlenecks that occur in bus-based architectures and provides the fastest possible connectivity between the CPUs, SDRAM (Synchronous DRAM) A type of dynamic RAM (DRAM) memory chip that has been widely used since the late 1990s. SDRAM chips eliminated wait states by dividing the chip into two cell blocks and interleaving data between them.  and I/O devices.

High Performance Busses Increase System Bandwidth

The integrated high speed bus interfaces, which includes DDR SDRAM, HyperTransport, SysAD and a local bus, provide low latency accesses into the main memory and high bandwidth to the I/O devices. The 200 MHz DDR (Double Data Rate) Refers to an SDRAM memory chip that increases performance by doubling the effective data rate of the frontside bus. For more details, see SDRAM.

DDR - Double Data Rate Random Access Memory
 memory controller offers 25.6 Gbit/s of memory bandwidth and has the flexibility to support both DDR SDRAM and the lower latency DDR FCRAM FCRAM Fast Cycle Random Access Memory
FCRAM Fast Cycle Ram
. Up to two Gigabytes of main memory can be addressed using 512 Mbit/s devices and the memory bus includes ECC (1) (Error-Correcting Code) A type of memory that corrects errors on the fly. See ECC memory.

(2) (Elliptic Curve Cryptography) A public key cryptography method that provides fast decryption and digital signature processing.
 protection. The HyperTransport I/O bus is new to PMC-Sierra's architecture and is based on an AMD (Advanced Micro Devices, Inc., Sunnyvale, CA, www.amd.com) A major manufacturer of semiconductor devices including x86-compatible CPUs, embedded processors, flash memories, programmable logic devices and networking chips.  standard. It is a 500 MHz DDR bus that delivers 16 Gbit/s of raw bus bandwidth for maximum performance. The

HyperTransport interface provides easy connection to a wide range of high speed networking peripherals. The SysAD bus is the same processor bus that is used on the RM7000(TM) and RM5200(TM) product families. The addition of a SysAD bus on the RM9000x2 enables connectivity to all SysAD-based peripheral devices and delivers a seamless upgrade path for existing RM7000 designs. The local bus provides a boot bus and connectivity to slower speed devices.

Additional RM9000x2 Features

The RM9000x2 provides very fast on-chip memory with 8 KB of integrated scratch RAM that is ideal for holding networking-related data structures such as descriptors and header information. A flexible, prioritized interrupt controller supports 256 HyperTransport interrupts, 10 external interrupts and interprocessor interrupts to support multiprocessing. A four-channel DMA (1) (Digital Media Adapter) See digital media hub.

(2) (Document Management Alliance) A specification that provides a common interface for accessing and searching document databases.
 controller provides high speed "any port-to-any port" data transfers between HyperTransport, SysAD, the main memory and integrated scratch RAM. The RM9000x2 supports EJTAG EJTAG Enhanced JTAG (MIPS processors)  debug To correct a problem in hardware or software. Debugging software means locating the errors in the source code (the program logic). Debugging hardware means finding errors in the circuit design (logical circuits) or in the physical interconnections of the circuits.  with debug modules integrated into each CPU. The EJTAG debug strategy ensures smooth debugging of both the hardware and software.

Availability

PMC-Sierra will begin sampling the RM9000x2 in Q4 of 2001. Evaluation boards will be available in Q1 of 2002.

About PMC-Sierra

PMC-Sierra is accelerating the broadband revolution. PMC-Sierra's extensive family of broadband communication semiconductors enables the equipment that makes up the backbone of the Internet. PMC-Sierra offers worldwide technical and sales support including a network of offices throughout North America, Europe and Asia. PMC-Sierra is publicly traded on the Nasdaq Stock Market Nasdaq stock market

The first electronic stock market listing over 5000 companies. The Nasdaq stock market comprises two separate markets, namely the Nasdaq National Market, which trades large, active securities and the Nasdaq Smallcap Market that trades emerging growth companies.
 under the symbol "PMCS". The company is included in the Nasdaq-100 Index (NDX NDX Index
NDX Index (File Name Extension)
NDX Northern Document Exchange
NDX Index File
) which contains the largest non-financial companies on the Nasdaq Stock Market. The Nasdaq-100 Index is the benchmark for the Nasdaq-100 Index Tracking Stock (Amex:QQQ QQQ

The Nasdaq-100 Index Tracking Stock. This is a tracking stock which trades like an index mutual fund which follows the Nasdaq 100 index. It trades continuously.


QQQ 
).

PMC-Sierra develops Internet Protocol (IP), ATM, SONET/SDH, T1/E1, T3/E3, Voice-over-Packet, wireless infrastructure, MIPS microprocessor, and Gigabit Ethernet solutions for wide area network (WAN), and Internet networking equipment. The company's quality system is registered with the Quality Management Institute to the ISO (1) See ISO speed.

(2) (International Organization for Standardization, Geneva, Switzerland, www.iso.ch) An organization that sets international standards, founded in 1946. The U.S. member body is ANSI.
 9001 standard. As co-founder of the SATURN(R) Development Group, PMC-Sierra works with over 30 other member companies to define and develop interoperable, standard-compliant solutions for high speed networking applications. For more information about PMC-Sierra, visit http://www.pmc-sierra.com.

Attachments:
Figure 1:  RM9000x2 Block Diagram
Figure 2:  RM9000x2 Applications
Technical Glossary

Technical Glossary
------------------
AMD              Advanced Micro Devices, Inc.
ATM              Asynchronous Transfer Mode
CPU              Central Processing Unit
DDR              Dual Data Rate
DMA              Direct Memory Access
ECC              Error Correcting Code
EJTAG            Enhanced Joint Test Action Group.  The adoption of a
                  single, industry-standard debugging interface for
                  the MIPS architecture
FCRAM            Fast Cycle Random Access Memory
Gbit/s           Billion bits per second
GHz              GigaHertz
I/O              Input/Output
IP               Internet Protocol
KB               KiloByte
Kbit             A thousand bits
MESI             Modified/Exclusive/Shared/Invalid
MHz              MegaHertz
MIPS             Millions of Instructions Per Second
MOESI            Modified/Owned (Modified-Shared)/Exclusive/Shared/
                  Invalid
RAM              Random Access Memory
RM9000x2(TM)     PMC-Sierra's RISCMark 9000x2(TM) MultiProcessor
                  Architecture
SATURN(R)        PMC-Sierra works with over 30 member companies to
                  define and develop interoperable, standard-compliant
                  solutions for high speed networking applications
SDH              Synchronous Digital Hierarchy
SDRAM            Synchronous Dynamic Random Access Memory
SONET            Synchronous Optical Network
SysAD            Multiplexed Address/Data System Bus
T1               North American standard for 1.544 Mbit/s transmission
T3               North American standard for a TDM
                  digital channel carrier that operates at
                  44.736 Mbit/s. It can multiplex 28 DS-1 signals and
                  is synonymous with DS-3
WAN              Wide Area Network


(c) Copyright PMC-Sierra, Inc. 2001. All rights reserved. SATURN(R)is a registered trademark of PMC-Sierra, Inc. RM9000x2(TM)Integrated Processor, RM7000(TM), RM5200(TM)and PMC-Sierra(TM)are trademarks of PMC-Sierra, Inc. HyperTransport(TM)is a trademark of AMD, Inc.

Note: A Graphic is available at URL URL
 in full Uniform Resource Locator

Address of a resource on the Internet. The resource can be any type of file stored on a server, such as a Web page, a text file, a graphics file, or an application program.
: http://www.businesswire.com/cgi-bin/photo.cgi?pw.061101/bb1 http://www.businesswire.com/cgi-bin/photo.cgi?pw.061101/bb1a
COPYRIGHT 2001 Business Wire
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Publication:Business Wire
Geographic Code:1CANA
Date:Jun 11, 2001
Words:1582
Previous Article:2Roam Brings Wireless Voice and Data Together With the Launch of ``2Roam Voice Services''; New Service To Provide Wireless Voice Functionality...
Next Article:14.3 GByte, Highest Capacity 3.5" SCSI Flash Drive Now Shipping from Adtron.
Topics:



Related Articles
V3 SEMICONDUCTOR DELIVERS LINUX SUPPORT FOR PMC-SIERRA'S MIPS PROCESSORS IN EMBEDDED PCI APPLICATIONS.(Company Business and Marketing)
PMC-Sierra's MIPS Processor Powers Best-Selling HP Laser Printer Family; 250MHz RM5261 Processor Selected for HP LaserJet 4100.
PMC-Sierra Ships Third Generation RM5200A MIPS Microprocessors.
PMC-Sierra Unveils Industry's Highest Performance MIPS Multiprocessor Architecture.
MIPS-Based Products Steal the Show at Embedded Processor Forum; NEC Electronics, PMC-Sierra Among Those Unveiling Products Based on 32-, 64-Bit MIPS...
API NetWorks Licenses HyperTransport Intellectual Property to PMC-Sierra for Development of Its Next Generation MIPS-based Processors.
PMC-SIERRA INTRODUCES 600 MHZ PIN-COMPATIBLE RM7000 64-BIT MIPS-BASED PROCESSORS.(PMC-Sierra RM7000C and PMC-Sierra RM7065C)(Product Announcement)
PMC-Sierra delivers SAS architectural solution for network attached and server attached storage.(PM8387 SXP, PM8389 SMC-L)
PMC-Sierra: Multi-Service Processor family for VoIP gateway apps.(voice over internet protocol)
PMC-Sierra demonstrates micro Multi-Service Provisioning Platform capabilities with new evaluation system.

Terms of use | Copyright © 2009 Farlex, Inc. | Feedback | For webmasters | Submit articles