为什么Python 中的代码速度较慢？(Why is this code slower in Python ?)

为什么Python 中的代码速度较慢？(Why is this code slower in Python ?)

我刚刚编写了这段代码，我想知道为什么它在Python 中表现得更差？ 所有平台都一样吗？ 这只是运气不好，还是Py通常较慢？

谢谢！

性能：

python 2.6 python .1 pypy 1.5 linux 2.2s 2.4s 0.8s os x 2.5s .4s 0.7s

代码:(对不起，它太仓促和低效！）

import itertools import random def fptp_draw(result): votes = [prefs[0] for prefs in result] counts = [len([v for v in votes if v == c]) for c in [1, 2, ]] s = sorted(counts) #print('fptp', counts) return s[-1] == s[-2] def av_remove(prefs, cand): if prefs[0] != cand: return prefs else: return prefs[1:] def av_draw(result): nv = len(result) cands = [1, 2, ] while True: votes = [prefs[0] for prefs in result] counts = [len([v for v in votes if v == c]) for c in cands] #print('av ', cands, counts) s = sorted(counts) if s[-1]*2 > nv: return False if len(cands) == 2: return True loser = cands[counts.index(s[0])] cands.remove(loser) result = [av_remove(prefs, loser) for prefs in result] return False #orders = list(itertools.permutati([1, 2, ])) orders = [[1, 2, ], [1, , 2], [2, 1, ], [2, , 1], [, 1, 2], [, 2, 1]] #results = list(itertools.product(*itertools.repeat(orders, 6))) #results = random.sample(results, 5) def rand_vote(): return [(orders) for i in range(1000)] n = fptp = av = 0 for j in range(1000): r = rand_vote() #print() #print(r) n += 1 if fptp_draw(r): fptp += 1 if av_draw(r): av += 1 print(fptp*100.0/n, av*100.0/n) print(n)

I just wrote this code, and I was wondering why it performs significantly worse in Python ? Is it the same across all platforms? Is this just bad luck, or is Py generally slower?

Thanks!

Performance:

python 2.6 python .1 pypy 1.5 linux 2.2s 2.4s 0.8s os x 2.5s .4s 0.7s

Code: (sorry it's so hasty and inefficient!)

import itertools import random def fptp_draw(result): votes = [prefs[0] for prefs in result] counts = [len([v for v in votes if v == c]) for c in [1, 2, ]] s = sorted(counts) #print('fptp', counts) return s[-1] == s[-2] def av_remove(prefs, cand): if prefs[0] != cand: return prefs else: return prefs[1:] def av_draw(result): nv = len(result) cands = [1, 2, ] while True: votes = [prefs[0] for prefs in result] counts = [len([v for v in votes if v == c]) for c in cands] #print('av ', cands, counts) s = sorted(counts) if s[-1]*2 > nv: return False if len(cands) == 2: return True loser = cands[counts.index(s[0])] cands.remove(loser) result = [av_remove(prefs, loser) for prefs in result] return False #orders = list(itertools.permutati([1, 2, ])) orders = [[1, 2, ], [1, , 2], [2, 1, ], [2, , 1], [, 1, 2], [, 2, 1]] #results = list(itertools.product(*itertools.repeat(orders, 6))) #results = random.sample(results, 5) def rand_vote(): return [(orders) for i in range(1000)] n = fptp = av = 0 for j in range(1000): r = rand_vote() #print() #print(r) n += 1 if fptp_draw(r): fptp += 1 if av_draw(r): av += 1 print(fptp*100.0/n, av*100.0/n) print(n)

最满意答案

Pyk通常比python 2.x 慢 。 这将随着时间的推移而改变，但pyk的重点是功能完整性和稳定性，而不是速度。

Pyk is generally slower than python 2.x. That will change as time passes, but the focus for pyk was feature completeness and stability, not speed.