// csa4.e   building block 4 bit by 4 bit csa multiplier
//          built in fad component

define fad(a, b, cin, s, cout)
circuits
  s    <= a ^ b ^ cin after 1ns;
  cout <= (a&b)|(a&cin)|(b&cin) after 1ns;
end circuits;
end fad;

define csa4(a[4], b[4], si[4], ci[4], pi[3], so[4], co[4], po[3])
  signal ce[4]; // temporary carry signals
  signal cf[4];
  signal cg[4];
  signal se[3];
  signal sf[3]; // temporary sum signals
  signal sg[4];
  signal ad[4];
  signal ae[4]; // output of 'a' mux's
  signal af[4];
  signal ag[4];
  signal zero[4] <= #h0;
circuits
  ad <= a when b[0] else zero after 1ns; // four 4 bit mux
  ae <= a when b[1] else zero after 1ns;
  af <= a when b[2] else zero after 1ns;
  ag <= a when b[3] else zero after 1ns;

  ad0 use fad(ad[0], si[0], ci[0], po[0], ce[0]); // first group 'd'
  ad1 use fad(ad[1], si[1], ci[1], se[0], ce[1]);
  ad2 use fad(ad[2], si[2], ci[2], se[1], ce[2]);
  ad3 use fad(ad[3], si[3], ci[3], se[2], ce[3]);

  ae0 use fad(ae[0], se[0], ce[0], po[1], cf[0]); // second group 'e'
  ae1 use fad(ae[1], se[1], ce[1], sf[0], cf[1]);
  ae2 use fad(ae[2], se[2], ce[2], sf[1], cf[2]);
  ae3 use fad(ae[3], pi[0], ce[3], sf[2], cf[3]);

  af0 use fad(af[0], sf[0], cf[0], po[2], cg[0]); // third group 'f'
  af1 use fad(af[1], sf[1], cf[1], sg[0], cg[1]);
  af2 use fad(af[2], sf[2], cf[2], sg[1], cg[2]);
  af3 use fad(af[3], pi[1], cf[3], sg[2], cg[3]);

  ag0 use fad(ag[0], sg[0], cg[0], so[0], co[0]); // last group 'g'
  ag1 use fad(ag[1], sg[1], cg[1], so[1], co[1]);
  ag2 use fad(ag[2], sg[2], cg[2], so[2], co[2]);
  ag3 use fad(ag[3], pi[2], cg[3], so[3], co[3]);
end circuits;
end csa4;