Polynomial Example

We will illustrate the command-response FIFO design mehodology with an example of a simple polynomial accelerator. The accelerator computes a polynomial function of x[k]:

y[k] = a[0] + a[1]*x[k] + ... + a[d-1]*x[k]**d

All the examples can be found in hwdesign/demos/stream with the main notebook in poly_demo.ipynb.

Protocol

The IP we will build will follow the folliwng command-response FIFO protocol:

  • The PS sends a PolyCmdHdr with the coefficients, coeffs, number of samples, nsamp, and a transaction ID, tx_id.
  • The IP repsonds with a PolyRespHdr message with the transcation ID, tx_id. This ensures that the IP is in sync and has received the data.
  • The PS then starts streaming the input data array x.
  • As the data streams in, the IP computes the output data array y.
  • When complete the IP sends a PolyRespFtr message with an indication of the number of samples and any error code.

Implementation

An implementation of the polynomial kernel and testbench can be found in the notebook. The implementation uses tthe Pysilicon package.
The notebook shows:

  • How to construct python PySilicon data schemas representing the messages
  • Auto-generate HLS C++ header files corresponding to the messages
  • Write Vitis HLS C++ code using the serialization and deserialization functions in the headers to transfer data between the kernel and IP using the messages
  • Test for synchronization errors
  • Run the C simulation and RTL simulation
  • Extract a VCD file from the RTL simulation
  • View the AXI4 timing bursts.
This site uses Just the Docs, a documentation theme for Jekyll.