Worst-case time complexity of Python's int.bit_length()

问题

When we call the function int.bit_length passing an integer n, is the worst-case time complexity O(log(n)) or Python uses some trick to improve it (e.g. storing the position of the most significant bit of n when it is created)?

回答1:

In CPython, for values with fewer internal-representation digits than PY_SSIZE_T_MAX/PyLong_SHIFT – i.e. fewer than PY_SSIZE_T_MAX binary digits – it’s calculated from the number of internal digits, yes:

msd = ((PyLongObject *)self)->ob_digit[ndigits-1];
msd_bits = bits_in_digit(msd);

if (ndigits <= PY_SSIZE_T_MAX/PyLong_SHIFT)
    return PyLong_FromSsize_t((ndigits-1)*PyLong_SHIFT + msd_bits);

Otherwise, it goes through bigints again, for overall time complexity of O(log log N) (which isn’t exactly true either in this strange mix of practice and theory, so…).

/* expression above may overflow; use Python integers instead */
result = (PyLongObject *)PyLong_FromSsize_t(ndigits - 1);
if (result == NULL)
    return NULL;
x = (PyLongObject *)PyLong_FromLong(PyLong_SHIFT);
if (x == NULL)
    goto error;
y = (PyLongObject *)long_mul(result, x);
Py_DECREF(x);
if (y == NULL)
    goto error;
Py_DECREF(result);
result = y;

x = (PyLongObject *)PyLong_FromLong((long)msd_bits);
if (x == NULL)
    goto error;
y = (PyLongObject *)long_add(result, x);
Py_DECREF(x);
if (y == NULL)
    goto error;
Py_DECREF(result);
result = y;

return (PyObject *)result;

tl;dr: it’s O(1)

来源：https://stackoverflow.com/questions/58461990/worst-case-time-complexity-of-pythons-int-bit-length

标签

python

time-complexity