Comments on: Fun with Fibonacci http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/ Blog of Will McGugan Fri, 21 Nov 2008 01:00:09 +0000 http://wordpress.org/?v=2.3.3 By: Wo kann ich filme downloaden? Danke http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-28936 Wo kann ich filme downloaden? Danke Tue, 01 Jul 2008 11:58:37 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-28936 <strong>coole 50 GB Erotik filme downloaden...</strong> Wo kann ich filme downloaden?... coole 50 GB Erotik filme downloaden…

Wo kann ich filme downloaden?…

]]> By: JeffH http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-12009 JeffH Thu, 13 Dec 2007 13:43:04 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-12009 Here is a non-recursive version <code> def fib(n): prev = -1 result = 1 for i in xrange(n+1): sum = result + prev prev = result result = sum return result </code> Time elapsed: 0.000813 Here is a non-recursive version

def fib(n):
prev = -1
result = 1
for i in xrange(n+1):
sum = result + prev
prev = result
result = sum
return result

Time elapsed: 0.000813

]]> By: eric casteleijn http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-12000 eric casteleijn Thu, 13 Dec 2007 11:22:02 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-12000 the first 2 lines of fib should read: if n == 0 or n == 1: return 1 the first 2 lines of fib should read:

if n == 0 or n == 1:
return 1

]]> By: Bob http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11996 Bob Thu, 13 Dec 2007 10:44:21 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11996 You could try something like this. import time def fib(n): # write Fibonacci series up to n """Print a Fibonacci series up to n.""" a, b = 0, 1 for i in range(n): #print b a, b = b, a+b startTime = time.clock() fib(36) print 'elapsed time: ', (time.clock() - startTime) You could try something like this.

import time

def fib(n): # write Fibonacci series up to n
“”"Print a Fibonacci series up to n.”"”
a, b = 0, 1
for i in range(n):
#print b
a, b = b, a+b

startTime = time.clock()

fib(36)

print ‘elapsed time: ‘, (time.clock() - startTime)

]]> By: filox http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11993 filox Thu, 13 Dec 2007 09:11:16 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11993 well, this didn't turn out as I'd hoped... for some reason, my python code was mangled :( well, this didn’t turn out as I’d hoped… for some reason, my python code was mangled :(

]]> By: filox http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11984 filox Thu, 13 Dec 2007 06:57:27 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11984 First, you have an error in the program - it says fib(2) = 1, which is wrong. Your whole sequence is one step behind. and since we're making fast algorithms, here's one even faster (doesn't spend as much memory either): from time import clock temp = [] temp.append(1) temp.append(1) def fib(n): global temp if n %d" % (i, fib(i)) print "Time elapsed: %f" % (clock() - start) First, you have an error in the program - it says fib(2) = 1, which is wrong. Your whole sequence is one step behind. and since we’re making fast algorithms, here’s one even faster (doesn’t spend as much memory either):

from time import clock
temp = []
temp.append(1)
temp.append(1)

def fib(n):
global temp
if n %d” % (i, fib(i))

print “Time elapsed: %f” % (clock() - start)

]]> By: JeffH http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11971 JeffH Thu, 13 Dec 2007 03:20:45 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11971 For comparison, running the code without the memoize decorator I get Time elapsed: 43.063779 Running Will's memoized version knocks it down to Time elapsed: 0.001341 Using shedskin, http://mark.dufour.googlepages.com/, on the non-memoized version runs in Time elapsed: 0.782071 (remember, I'm not trading memory for speed) Making a shared library out of fib with shedskin, then importing it and using it results in: Time elapsed: 1.242797 You really need to check out shedskin -- c++ speeds, python code to maintain For comparison, running the code without the memoize decorator I get
Time elapsed: 43.063779

Running Will’s memoized version knocks it down to
Time elapsed: 0.001341

Using shedskin, http://mark.dufour.googlepages.com/, on the non-memoized version runs in
Time elapsed: 0.782071
(remember, I’m not trading memory for speed)

Making a shared library out of fib with shedskin, then importing it and using it results in:
Time elapsed: 1.242797

You really need to check out shedskin — c++ speeds, python code to maintain

]]> By: billg http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11953 billg Wed, 12 Dec 2007 21:00:16 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11953 should be even faster with xrange instead of range (unless you're using python3000). of course, you probably won't be able to measure the speedup. should be even faster with xrange instead of range (unless you’re using python3000). of course, you probably won’t be able to measure the speedup.

]]> By: Bob http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11946 Bob Wed, 12 Dec 2007 19:06:14 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11946 I just took my own advice that I just gave you. I ran on WIndows XP with time.time() and I also got 0 seconds. When I ran using time.clock() I got about 3 milliseconds. I just took my own advice that I just gave you. I ran on WIndows XP with time.time() and I also got 0 seconds. When I ran using time.clock() I got about 3 milliseconds.

]]> By: Bob http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11945 Bob Wed, 12 Dec 2007 19:00:28 +0000 http://www.willmcgugan.com/2007/12/12/fun-with-fibonacci/#comment-11945 You must be using some form of Windows OS. Python time.time() resolution on Windows XP is around 15 milliseconds. So you have shown that your algorithm runs in less than about 15 ms on your machine. Try it again using time.clock(). Python time.clock() on my Windows XP box has about 2 microsecond resolution. You must be using some form of Windows OS. Python time.time() resolution on Windows XP is around 15 milliseconds. So you have shown that your algorithm runs in less than about 15 ms on your machine. Try it again using time.clock(). Python time.clock() on my Windows XP box has about 2 microsecond resolution.

]]>