comparing_speeds.py

import timeit
from numpy import mean

nqueens_gen_setup = """def threatened(pos, past, n):
    for i in range(len(past)):
        if past[i] == pos or abs(pos - past[i]) == len(past) - i:
            return True
    return False

def nqueens(n, past=None):
    if past == None:
        for i in range(n):
            yield from nqueens(n, [i+1])
    else:
        if len(past) == n:
            yield past[:]
        else:
            for pos in range(1,n+1):
                if not threatened(pos, past, n):
                    yield from nqueens(n,past + [pos])

list(nqueens(8))"""

nqueens_global_setup = """
solutions = []

def threatened(pos, past, n):
    for i in range(len(past)):
        if past[i] == pos or abs(pos - past[i]) == len(past) - i:
            return True
    return False

def nqueens(n, past=None):
    if past == None:
        return [nqueens(n, [i+1]) for i in range(n)]
    if len(past) == n:
        solutions.append(past)
    else:
        [nqueens(n, past + [pos]) for pos in range(1, n+1) if not threatened(pos, past, n)]

nqueens(8)
"""

print(mean(timeit.Timer(setup=nqueens_gen_setup).repeat(20, 1)))
print(mean(timeit.Timer(setup=nqueens_global_setup).repeat(20,1)))

## I played with the numbers of repeats, nqueens_gen_setup is slightly faster every time. But I bet I could cut the speed on 
## nqueens_gobal_setup to be more competitive!

nqueens.py

solutions = []

def threatened(pos, past, n):
    for i in range(len(past)):
        if past[i] == pos or abs(pos - past[i]) == len(past) - i:
            return True
    return False

def nqueens(n, past=None):
    if past == None:
        return [nqueens(n, [i+1]) for i in range(n)]
    if len(past) == n:
        solutions.append(past)
    else:
        [nqueens(n, past + [pos]) for pos in range(1, n+1) if not threatened(pos, past, n)]

nqueens(8)
print(len(solutions))