module mult8x8(y,a,b,clock);
input [8:0] a,b;
output [16:0] y;
input clock;

reg [8:0] sum,dif,alb,bla;
reg sign0;
always @(posedge clock) begin : stage0
  sum <= a[7:0] + b[7:0];  // add the magnitudes
  // have to deal with underfow in the difference
  alb = a[7:0] - b[7:0];
  bla = b[7:0] - a[7:0];
  if (alb[8]) dif <= bla;
  else dif <= alb;
  // compute sign
  sign0 <= a[8] ^ b[8];         
end

reg [17:0] sum_squared, dif_squared; 
reg sign1;
always @(posedge clock) begin : stage1
  sign1 <= sign0;
  sum_squared <= sum*sum;
  dif_squared <= dif*dif;
end

reg [17:0] magnitude;
reg [16:0] y;
always @(posedge clock) begin : stage2
  magnitude = sum_squared - dif_squared;
  y <= magnitude >> 2;
  y[16] <= sign1;
end

endmodule