Harold's comment is correct. Consider what happens for float inputs like 5000000000 * 1.0 . Conversion to int32_t with cvtps2dq will give ...
確定! 回上一頁