Comments on: The Sieve of Atkin in C# http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/ With a hint of Social Ineptitude Tue, 07 Feb 2012 09:05:01 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: To find the list of primes till N. | I Live to Seek… http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-9943 To find the list of primes till N. | I Live to Seek… Wed, 26 Oct 2011 02:36:16 +0000 http://www.geekality.net/?p=674#comment-9943 [...] Wikipedia Implementation C# [...] [...] Wikipedia Implementation C# [...]

]]> By: Ben http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-6260 Ben Wed, 22 Jun 2011 13:37:31 +0000 http://www.geekality.net/?p=674#comment-6260 The list of primes takes more memory stored as a list than as the bool array it's calculated in. It also requires memory for both to change from one to the other. I replaced primes with the isPrime array (moved to class-level), changed the IEnumerator to this: <code> public IEnumerator<ulong> GetEnumerator() { if (!primeFlag) { FindPrimes(); isPrime[2] = true; isPrime[3] = true; primeFlag = true; } for (ulong i = 2;i<limit;i++) { if (isPrime[i]){ yield return i; } } } </code> If memory still remains a problem, there are a number of memory optimisations that can be made, some at the expense of performance (for example, only representing odd numbers in isPrime). The list of primes takes more memory stored as a list than as the bool array it’s calculated in.

It also requires memory for both to change from one to the other.

I replaced primes with the isPrime array (moved to class-level), changed the IEnumerator to this:

public IEnumerator<ulong> GetEnumerator() {
            if (!primeFlag) {
                FindPrimes();
                isPrime[2] = true;
                isPrime[3] = true;
                primeFlag = true;
            }
            for (ulong i = 2;i<limit;i++) {
                if (isPrime[i]){
                    yield return i;
                }
            }
        }

If memory still remains a problem, there are a number of memory optimisations that can be made, some at the expense of performance (for example, only representing odd numbers in isPrime).

]]> By: Torleif http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-2647 Torleif Fri, 03 Dec 2010 20:01:32 +0000 http://www.geekality.net/?p=674#comment-2647 Like I said, I am not going to try to explain why the algorithm works. But it does, and there are tests in my code that proves it. Just try it out for yourself :) You can read about the algorithm <a href="http://en.wikipedia.org/wiki/Sieve_of_Atkin#Algorithm" rel="nofollow">at wikipedia</a>, Like I said, I am not going to try to explain why the algorithm works. But it does, and there are tests in my code that proves it. Just try it out for yourself :)

You can read about the algorithm at wikipedia,

]]> By: Ajay http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-2638 Ajay Fri, 03 Dec 2010 15:38:27 +0000 http://www.geekality.net/?p=674#comment-2638 I still dnt understand. how could this code works?? i mean take an example of nubmer 65.. 4*2*2 + 7*7 =65..... markdes as true in first loop.... after then.... it would never be unmarked and would be in output inspite of being a composite.... admin plzz solve this issue,,,, I still dnt understand. how could this code works??
i mean take an example of nubmer 65..

4*2*2 + 7*7 =65….. markdes as true in first loop….
after then…. it would never be unmarked and would be in output inspite of being a composite….

admin plzz solve this issue,,,,

]]> By: yofreke http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-1689 yofreke Tue, 26 Oct 2010 06:16:15 +0000 http://www.geekality.net/?p=674#comment-1689 (my code uses longs) long xy = x * x + y * y; var n = 4 * xy; ... n = 3 * xy; I tried this and it took about 1/3 less of the time with my setup. (my code uses longs)
long xy = x * x + y * y;
var n = 4 * xy;
…
n = 3 * xy;
I tried this and it took about 1/3 less of the time with my setup.

]]> By: Torleif http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-1581 Torleif Wed, 20 Oct 2010 00:44:04 +0000 http://www.geekality.net/?p=674#comment-1581 Well, I suppose, but I have a feeling that allocating the variables might be slower than just doing the calculation... have you tried it out? Well, I suppose, but I have a feeling that allocating the variables might be slower than just doing the calculation… have you tried it out?

]]> By: Torleif http://www.geekality.net/2009/10/19/the-sieve-of-atkin-in-c/#comment-1579 Torleif Tue, 19 Oct 2010 22:02:17 +0000 http://www.geekality.net/?p=674#comment-1579 Ooh, good catch. Hadn't tried it with that large values. Seems like it will fail for two reasons. First of all, the array can only contain <code inline="true">int32.MaxValue</code> items. Secondly, you'll run out of memory :P Do you have any suggestions on how to fix it? I have a feeling this more a flaw of the algorithm itself than my implementation though... but yeah, if you have any smart solutions please let me know! Ooh, good catch. Hadn’t tried it with that large values. Seems like it will fail for two reasons. First of all, the array can only contain int32.MaxValue items. Secondly, you’ll run out of memory :P

Do you have any suggestions on how to fix it? I have a feeling this more a flaw of the algorithm itself than my implementation though… but yeah, if you have any smart solutions please let me know!

]]>