vnclipu.wx

{
  let SEW      = get_sew();
  let LMUL_pow = get_lmul_pow();
  let num_elem = get_num_elem(LMUL_pow, SEW);
  let SEW_widen      = SEW * 2;
  let LMUL_pow_widen = LMUL_pow + 1;

  if  illegal_variable_width(vd, vm, SEW_widen, LMUL_pow_widen) |
      not(valid_reg_overlap(vs2, vd, LMUL_pow_widen, LMUL_pow))
  then { handle_illegal(); return RETIRE_FAIL };

  let 'n = num_elem;
  let 'm = SEW;
  let 'o = SEW_widen;

  let vm_val  : vector('n, dec, bool)     = read_vmask(num_elem, vm, 0b00000);
  let vd_val  : vector('n, dec, bits('m)) = read_vreg(num_elem, SEW, LMUL_pow, vd);
  let rs1_val : bits('m)                  = get_scalar(rs1, SEW);
  let vs2_val : vector('n, dec, bits('o)) = read_vreg(num_elem, SEW_widen, LMUL_pow_widen, vs2);
  result      : vector('n, dec, bits('m)) = undefined;
  mask        : vector('n, dec, bool)     = undefined;

  (result, mask) = init_masked_result(num_elem, SEW, LMUL_pow, vd_val, vm_val);

  assert(SEW_widen <= 64);
  foreach (i from 0 to (num_elem - 1)) {
    if mask[i] then {
      let shift_amount = get_shift_amount(rs1_val, SEW_widen);
      let rounding_incr = get_fixed_rounding_incr(vs2_val[i], shift_amount);
      result[i] = match funct6 {
        NX_VNCLIPU => {
                        let result_wide = (vs2_val[i] >> shift_amount) + zero_extend('o, rounding_incr);
                        unsigned_saturation('m, result_wide)
                      },
        NX_VNCLIP  => {
                        let v_double : bits('m * 4) = sign_extend(vs2_val[i]);
                        let result_wide = slice(v_double >> shift_amount, 0, 'o) + zero_extend('o, rounding_incr);
                        signed_saturation('m, result_wide)
                      }
      }
    }
  };

  write_vreg(num_elem, SEW, LMUL_pow, vd, result);
  vstart = zeros();
  RETIRE_SUCCESS
}