了解递归调用中的Python列表内存使用情况

Question

I just wanted to check if I understood Python's memory management correctly.

The following function would use O(j) memory, but not O(nj) memory, since the parameter n is a reference to the list, but not the list itself.

def cool(n,j):
    if j == 0:
        return
    return cool(n,j-1)

Also, say this function was written in C, am I correct in assuming that, with C, its memory usage would be O(nj) , since a copy of the list/array would be created for each recursive call.

Answer 1

Yes, that is correct. Each recursive step will use O(1) memory: function call overhead, and one pointer for each function parameter.

You can experimentally verify this by creating a large object ( x = "x" * 1024 * 1024 * 100 ), then recursing on it a number of times and checking the process's memory usage:

def recurse(x, n):
    if n == 0:
        raw_input("done! Check memory usage now.")
        return
    return recurse(x, n - 1)

recurse(x, 1000)