+ +
- -
Systèmes d'Exploitation
Calendrier  Détails
Programmation
 
Calendrier  Détails
Processeurs ARM
 
Calendrier  Détails
Processeurs PowerPC
 
 
 
 
Calendrier  Détails
Communications
 
 
Calendrier  Détails
+ +
> >
- -

 
ac6 >> ac6-training >> Processeurs PowerPC >> IBM processors >> CoreConnect Télécharger le catalogue Télécharger la page Ecrivez nous Version imprimable

P4 CoreConnect

This course covers the CoreConnect specification, explaining PLB, OPB, DCR buses and bridges

formateur
Objectives
  • The course describes the 3 buses specified by the IBM CoreConnect specification : PLB, OPB and DCR.
  • It explains also the operation of bus bridges PLB-to-OPB and OPB-to-PLB.
  • All parameters of the Xilinx CoreConnect infrastructure logicores are described in detail.
  • Labs have been developed to become familiar with the simulation toolkit : Bus Functional Models (BFM) and CTG (CoreConnect Test Generator).
  • The course focuses on bus error recovery through syndrome registers.
  • 128-bit PLB, also known as PLB4, is fully covered including 2-way crossbar implementation.
  • The course explains how to tune programmable parameters through the PLB performance monitor.

    *

  • This training has been delivered several times to engineers developing ASICs based on Power cores and to engineers developing SoCs based on Xilinx FPGAs containing Power cores.
A more detailed course description is available on request at training@ac6-training.com
Prerequisites
  • Experience of a parallel digital bus is mandatory.

INTRODUCTION TO CoreConnect
  • SoC organization
  • Intellectual Property reuse by using common bus for inter-macro communication
  • The IBM 3-bus for interconnecting cores : PLB, OPB and DCR
  • Benefits of DCR compared to memory-mapped IOs
  • The infrastructure cores developped by Xilinx
THE PLB
  • · Arbitration
  • · Bus time-out detection
  • · Locked transfer
  • · Address pipelining capability
  • · Differences between a 1-deep and a N-deep (N>2) pipeline implementation
  • Single data, burst and line transfer timing diagrams
  • Read burst and write burst terminations
  • Dynamic bus width adaptation
  • PLB usage in Xilinx FPGAs
  • The PLB Xilinx logicore
FIXING BUS ERRORS
  • Parity generation and checking
  • Slave error report to masters
  • Syndrome registers
THE PLB PERFORMANCE MONITOR
  • Use of the PPM to tune programmable parameters
  • Event counting, duration measurement
  • Connection of the PPM to the PLB fabric
  • Pipeline stage usage tracking
PLB ARBITRATION
  • Central arbitration mechanism
  • Fixed and rotative priority schemes
  • PLB watchdog timer
  • Programming interface
  • Xilinx PLB arbiter operating modes
THE 128-BIT 2-WAY CROSSBAR
  • Concurrent read transactions and concurrent write transactions
  • Highlighting address path, read data path and write data path
  • Selecting the slave bus segment, PCBC register programming
THE OPB
  • · Dynamic bus sizing vs Byte Enables
  • · Distributed multiplexing
  • Arbitration
  • OPB interface for master, slave, arbiter and DMAs
  • Slave retry
  • Logicore Xilinx OPB with OPB arbiter
  • Connection to OPB through IPIF
THE PLB-to-OPB BRIDGE
  • Block diagram and data flows
  • Internal data buffers structure
  • PLB-to-OPB signals
  • Bridge control registers
  • Xilinx PLB-to-OPB bridge user configurable parameters
  • Definition of address ranges allowing PLB masters to access the OPB bus
THE OPB-to-PLB BRIDGE
  • Block diagram and data flows
  • Internal data buffers structure
  • Synchronization with the PLB-to-OPB bridge
  • Bridge control registers
  • Xilinx OPB-to-PLB bridge user configurable parameters
  • Definition of address ranges allowing OPB masters to access the PLB bus
THE DCR BUS
  • Features
  • Bus operation : bypass mux use
  • The DCR Xilinx logicore
SIMULATING CoreConnect BUSES
  • Description of the simulation tools provided by IBM Microelectronics : BFMs and CTG
  • Step-by-step explanation of the simulation flow
  • Development of a testbench to test a PLB IP