GTS AXI Streaming Intel® FPGA IP for PCI Express* User Guide

ID 813754
Date 8/07/2024
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

4.8.5. Data and Header Packing Scheme

GTS AXI Streaming IP uses simple packing scheme with an inline header packet format where the header is presented on the same bus as data.

The simple packing scheme inserts a header starting at a fixed location. The header always starts from Byte Index 0. The rule of header starting on Byte Index 0 constrains the design to send one packet per cycle.

The PCIe* 4.0 x2/ PCIe* 4.0 x1 and all PCIe* 3.0 modes except PCIe* 3.0 x8 can take advantage of the 128-bit wide PLD interface. The 128-bit wide interface allows you to build design with smaller data path width. With the 128-bit interface, the header is transferred over two clock cycles provided tready signal is asserted.

The HIP IF Adaptor Tdata bus width varies based on link width and link speed to meet specific bandwidth requirement.

The following table lists the simple packing data width and the optimum PLD clock requirements to meet the link bandwidth goal.
Table 20.  Simple Packing Data Width and Optimum PLD Clock Frequency
Mode Data Width PLD Clock Frequency (MHz) Number of Streams
PCIe* 4.0 x8 512 500 1
PCIe* 4.0 x4/ PCIe* 3.0 x8 256 350 1
PCIe* 4.0 x2/ PCIe* 3.0 x4 128 300 1
PCIe* 4.0 x1/ PCIe* 3.0 x2/ PCIe* 3.0 x1 128 200 1