`timescale 1ns / 1ps
//////////////////////////////////////////////////////////////////////////////////
// Company: 
// Engineer: 
// 
// Create Date:    00:07:27 10/07/2007 
// Design Name: 
// Module Name:    fir_pipline_lms 
// Project Name: 
// Target Devices: 
// Tool versions: 
// Description: 
//
// Dependencies: 
//
// Revision: 
// Revision 0.01 - File Created
// Additional Comments: 
//
//////////////////////////////////////////////////////////////////////////////////
module fir_pipeline_lms (clk, x_in, d_in, e_out, y_out);

  parameter W1 = 16,   //輸入信號的位寬
            W2 = 32;  //乘法器的輸出位寬

  input clk;  //輸入時鐘
  input [W1-1:0] x_in, d_in;  //輸入信號和參考信號 
  output [W2-1:0] e_out, y_out;  //誤差信號和輸出信號

  //聲明中間變量
  reg  [W1-1:0] x, x0, x1, x2, x3, x4, f0, f1;  
  reg  [W1-1:0] d0, d1, d2, d3; 
  wire [W1-1:0] emu;
  wire [W2-1:0] p0, p1, xemu0, xemu1; //紀錄乘積
  wire  [W2-1:0]  y, sxty, e, sxtd; 

  //擴展輸入的參考信號
  assign  sxtd = {{16{d3[7]}}, d3};
  
  //存儲數據和系數. 4級移位，補償乘法器的流水線所帶來的延遲
  always @(posedge clk) begin
      d0 <= d_in; //構成移位寄存器
      d1 <= d0;
      d2 <= d1;
      d3 <= d2;
      x0 <= x_in; 
      x1 <= x0;   
      x2 <= x1;
      x3 <= x2;
      x4 <= x3;
      f0 <= f0 + xemu0[31:16]; //相當于步長因子為1/2
      f1 <= f1 + xemu1[31:16]; 
  end

  //調用IPCore生成乘法器lmsmult時，選了最大的流水線級數3級
  //計算p(i) = f(i) * x(i); 
  // x0*f0 = p0  
  lmsmult mul1(
   .clk(clk), .a(x0), .b(f0), .q(p0));    
    // x1*f1 = p1 
  lmsmult mul2(              
   .clk(clk), .a(x1), .b(f1), .q(p1));
  
  // 計算濾波器的輸出
  assign y = p0 + p1;  
  
  assign  sxty = {{15{y[15]}}, y[31:16]};

  assign e = sxtd - sxty;
  assign emu = e[16:1]; 
  
  //計算xemu(i) = emu * x(i);
  //xemu0 = emu * x0;
  lmsmult mul3(
	.clk(clk), .a(x3), .b(emu), .q(xemu0));
  //xemu1 = emu * x1;  
  lmsmult mul4(            
   .clk(clk), .a(x4), .b(emu), .q(xemu1));

  assign  y_out  = y;   
  assign  e_out  = e;

endmodule