PLX to detail BulletTrain on-chip architecture
![]() |
PLX to detail BulletTrain on-chip architecture
By Loring Wirbel, EE Times
January 15, 2003 (10:06 a.m. EST)
URL: http://www.eetimes.com/story/OEG20030115S0015
SAN JOSE, Calif. PLX Technology Inc. will disclose architectural details next week of a generic on-chip fabric architecture called BulletTrain, which represents the company's bid to offer interconnect products that meet the PCI Express Base specification and PCI Express Advanced Switching specification.
At the Bus and Board Conference, which begins Jan. 20 in Long Beach, Calif., PLX will explain how the core SwitchYard fabric at the heart of its BulletTrain architecture can be linked to expandable ingress and egress ports, making it easier to broaden applications than with variable-sized buffers.
Each port that connects to a specific I/O layer is defined by what PLX calls "Station IP," which refers to generic intellectual property rather than Internet Protocol links. If a company has its own IP cores, PLX allows them to link to BulletTrain through software "gaskets" defined through application programming interfaces. Product ma nager Danny Chi said that defining a rule-based architecture appropriate for any PCI Express implementation eliminated the need for different switches and protocol bridges for different designs. The SwitchYard fabric is a point-to-point mesh that can expand in multiple dimensions, and the BulletTrain topology with SwitchYard at its center can be used with alternative interconnects such as HyperTransport.
The only thing that changes with more ingress and egress points is the number of packet queues, Chi said. Since the fabric is intelligent, no processor interdiction from the control plane is required. PLX will offer its first PCI Express base chips utilizing BulletTrain in the second half of 2003, followed by PCI Express Advanced Switching chips in early 2004.
Related News
- Arm's solution to the future needs of AI, security and specialized computing is v9
- RISC-V International Unveils Fast Track Architecture Extension Process and Ratifies ZiHintPause Extension
- Think Silicon to introduce a new Inference Micro GPU Architecture based on RISC-V at Linley Fall Virtual Processor Conference
- The Industry's First SoC FPGA Development Kit Based on the RISC-V Instruction Set Architecture is Now Available
- SiFive To Introduce New RISC-V Processor Architecture and RISC-V PC at Linley Fall Virtual Processor Conference
Breaking News
- Hardent Selected as Design Services Provider for New Xilinx Kria SOMs
- Codasip Announces FPGA Evaluation Platforms for RISC-V Processor Cores
- Renesas and SiFive Partner to Jointly-Develop Next-Generation High-End RISC-V Solutions for Automotive Applications
- Synopsys Introduces PrimeLib Unified Library Characterization and Validation Solution for Accelerated Access to Advanced Process Nodes
- CAES Gaisler Signs Contract with the European Space Agency for New Advanced Space Processor
Most Popular
- TSMC certifies Aprisa place-and-route solution from Siemens on TSMC's N6 process
- TSMC Boosts Capital Budget Again, to $30B
- Expedera Introduces Its Origin Neural Engine IP With Unrivaled Energy-Efficiency and Performance
- Revenue of Top 10 IC Design (Fabless) Companies for 2020 Undergoes 26.4% Increase YoY Due to High Demand for Notebooks and Networking Products, Says TrendForce
- Groq Closes $300 Million Fundraise
![]() |
E-mail This Article | ![]() |
![]() |
Printer-Friendly Page |