目录
目录README.md

PyHCL

Build Status codecov PyPI

PyHCL is a hardware construct language like Chisel but more lightweight and more relaxed to use. As a novel hardware construction framework embedded in Python, PyHCL supports several useful features include object-oriented, functional programming, and dynamically typed objects.

The goal of PyHCL is providing a complete design and verification tool flow for heterogeneous computing systems flexibly using the same design methodology.

PyHCL is powered by FIRRTL, an intermediate representation for digital circuit design.

PyHCL-generated circuits can be compiled to the widely-used HDL Verilog.

Attention: The back end of the compilation is highly experimental.

Getting Started

Writing A Full Adder

PyHCL defines modules using only simple Python syntax that looks like this:

from pyhcl import *

class FullAdder(Module):
    io = IO(
        a=Input(Bool),
        b=Input(Bool),
        cin=Input(Bool),
        sum=Output(Bool),
        cout=Output(Bool),
    )

    # Generate the sum
    io.sum @= io.a ^ io.b ^ io.cin

    # Generate the carry
    io.cout @= io.a & io.b | io.b & io.cin | io.a & io.cin

Compiling To High FIRRTL

Compiling module by calling compile_to_highform:

Emitter.dump(Emitter.emit(FullAdder(), HighForm), "FullAdder.fir")

Will generate the following FIRRTL codes:

circuit FullAdder :
  module FullAdder :
    input clock : Clock
    input reset : UInt<1>
    output io : {flip a : UInt<1>, flip b : UInt<1>, flip cin : UInt<1>, s : UInt<1>, cout : UInt<1>}
  
    node _T = xor(io.a, io.b)
    node _T_1 = xor(_T, io.cin)
    io.s <= _T_1
    node _T_2 = and(io.a, io.b)
    node _T_3 = and(io.a, io.cin)
    node _T_4 = or(_T_2, _T_3)
    node _T_5 = and(io.b, io.cin)
    node _T_6 = or(_T_4, _T_5)
    io.cout <= _T_6

Compiling To Lowered FIRRTL

Compiling module by calling compile_to_lowform:

Emitter.dump(Emitter.emit(FullAdder(), LowForm), "FullAdder.lo.fir")

Will generate the following FIRRTL codes:

circuit FullAdder :
  module FullAdder :
    input clock : Clock
    input reset : UInt<1>
    input io_a : UInt<1>
    input io_b : UInt<1>
    input io_cin : UInt<1>
    output io_s : UInt<1>
    output io_cout : UInt<1>
  
    node _T = xor(io_a, io_b)
    node _T_1 = xor(_T, io_cin)
    io_s <= _T_1
    node _T_2 = and(io_a, io_b)
    node _T_3 = and(io_a, io_cin)
    node _T_4 = or(_T_2, _T_3)
    node _T_5 = and(io_b, io_cin)
    node _T_6 = or(_T_4, _T_5)
    io_cout <= _T_6

Compiling To Verilog

Compiling module by calling compile_to_verilog:

Emitter.dump(Emitter.emit(FullAdder(), Verilog), "FullAdder.v")

Then FullAdder.v will be generated:

module FullAdder(
  input        clock,
  input        reset,
  input        io_a,
  input        io_b,
  input        io_cin,
  output        io_s,
  output        io_cout
);

  assign io_s = io_a ^ io_b ^ io_cin;
  assign io_cout = io_a & io_b | io_a & io_cin | io_b & io_cin;
  
endmodule

Features

  • Supports multiple data types: UInt, SInt, Vector, Bundle, Clock, Memory, and casual combination between them.
  • Supports object-oriented inheritance, can compose modules by writing fewer codes.
  • Supports a bunch of convenient operations, such as the addition of UInts, SInts, Vectors and Bundles.
  • Supports the parameterization of variables, such as bit width, with the syntax facilities of the host language Python.

TODO

  • Supports more operations
  • PyHCL’s verification facility
邀请码
    Gitlink(确实开源)
  • 加入我们
  • 官网邮箱:gitlink@ccf.org.cn
  • QQ群
  • QQ群
  • 公众号
  • 公众号

©Copyright 2023 CCF 开源发展委员会
Powered by Trustie& IntelliDE 京ICP备13000930号