Intel 41210 User Manual
Intel
®
41210 Serial to Parallel PCI Bridge Developer’s Manual
11
Introduction
Introduction
1
The Intel
®
41210 Serial to Parallel PCI Bridge (also called the 41210 Bridge or the 41210)
integrates two PCI Express*-to-PCI/PCI-X bridges. Each bridge follows the PCI-to-PCI Bridge
programming model. The PCI Express* port is compatible with the PCI Express* Specification,
Revision 1.0a. The two PCI bus interfaces are comparable with the PCI Local Bus Specification,
Revision 2.3 and the PCI-X Addendum to the PCI Local Bus Specification, Revision 1.0b.
programming model. The PCI Express* port is compatible with the PCI Express* Specification,
Revision 1.0a. The two PCI bus interfaces are comparable with the PCI Local Bus Specification,
Revision 2.3 and the PCI-X Addendum to the PCI Local Bus Specification, Revision 1.0b.
1.1
PCI Express* Interface Features
•
PCI Express* Specification, Revision 1.0a
•
Support for single ×8, single ×4 or single x1 PCI Express* operation
•
64-bit addressing support
•
32-bit CRC (cyclic redundancy checking) covering all transmitted data packets
•
16-bit CRC on all link message information
•
Raw bit-rate on the data pins of 2.5 Gbit/s, resulting in a raw bandwidth per pin of 250 MB/s
•
Maximum realized bandwidth on PCI Express* interface of 2 GB/s (in ×8 mode) in each
direction simultaneously, for an aggregate of 4 GB/s
direction simultaneously, for an aggregate of 4 GB/s
1.2
PCI-X Interface Features
•
PCI Local Bus Specification, Revision 2.3
•
PCI-to-PCI Bridge Specification, Revision 1.1
•
PCI-X Addendum to the PCI Local Bus Specification, Revision 1.0b
•
64-bit 66 MHz, 3.3 V, not 5 V tolerant
•
On-Die Termination (ODT) with 8.3 K
Ω
pull-up to 3.3 V for PCI signals
•
Six external REQ/GNT pairs for internal arbiter on segment A and B respectively
•
Programmable bus parking on either the last agent or always on Intel
®
41210 Serial to Parallel
PCI Bridge
•
Two-level programmable round-robin internal arbiter with Multi-Transaction Timer (MTT)
•
External PCI clock-feed support for asynchronous primary and secondary domain operation
•
64-bit addressing for upstream and downstream transactions
•
Downstream LOCK# support
•
No upstream LOCK# support
•
PCI fast Back-to-Back capable as target
•
Up to four active and four pending upstream memory read transactions