#ifdef TFM_SQR4
void fp_sqr_comba4(fp_int *A, fp_int *B)
{
   fp_digit *a, b[8], c0, c1, c2, sc0, sc1, sc2;

   a = A->dp;
   COMBA_START; 

   /* clear carries */
   CLEAR_CARRY;

   /* output 0 */
   SQRADD(a[0],a[0]);
   COMBA_STORE(b[0]);

   /* output 1 */
   CARRY_FORWARD;
   SQRADD2(a[0], a[1]); 
   COMBA_STORE(b[1]);

   /* output 2 */
   CARRY_FORWARD;
   SQRADD2(a[0], a[2]); SQRADD(a[1], a[1]); 
   COMBA_STORE(b[2]);

   /* output 3 */
   CARRY_FORWARD;
   SQRADD2(a[0], a[3]); SQRADD2(a[1], a[2]); 
   COMBA_STORE(b[3]);

   /* output 4 */
   CARRY_FORWARD;
   SQRADD2(a[1], a[3]); SQRADD(a[2], a[2]); 
   COMBA_STORE(b[4]);

   /* output 5 */
   CARRY_FORWARD;
   SQRADD2(a[2], a[3]); 
   COMBA_STORE(b[5]);

   /* output 6 */
   CARRY_FORWARD;
   SQRADD(a[3], a[3]); 
   COMBA_STORE(b[6]);
   COMBA_STORE2(b[7]);
   COMBA_FINI;

   B->used = 8;
   B->sign = FP_ZPOS;
   memcpy(B->dp, b, 8 * sizeof(fp_digit));
   fp_clamp(B);
}
#endif