Intel® FPGA AI Suite: IP Reference Manual

ID 768974
Date 9/06/2023
Public

A newer version of this document is available.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

Document Table of Contents

3. Intel® FPGA AI Suite IP Generation Utility

The Intel® FPGA AI Suite IP generation utility reads an input Architecture Description File (.arch) and places generated IP into an IP library that can be imported into Platform Designer or used directly in a pure RTL design. The generated IP is configured based on the Architecture Description (.arch) File, which defines the following items:

  • Top RTL parameters are set to reflect values in the Architecture Description File
  • The Discovery ROM MIF is created to specify an architecture hash and Intel® FPGA AI Suite compiler version string

The generation utility also copies all the required RTL files into a single directory. As convenience options, the utility can also help to create .qsf files and wrappers for the PCIe Example Design or to compile an instance of the IP with a dummy BSP interface.

The IP generation utility generates either an unlicensed or a licensed copy of the IP:
Unlicensed IP
The unlicensed IP has a limit of 10000 inferences. After 10000 inferences, the unlicensed IP refuses to perform any additional inference and a bit in the CSR is set. For details about the CSR bit, refer to DMA Descriptor Queue.
Licensed IP
The licensed IP has no inference limitation
The IP generation utility checks for an Intel® FPGA AI Suite IP license before generating the IP. The utility prints messages to stdout that show the license status.

You can use either licensed and unlicensed IP for bitstream generation so that you can fully test your design during the evaluation process.