以多进程方式计算斐波那契数？

Question

I am writing multi process fibonacci number calculator , I have a file that keeps track of the fibonacci numbers , first process open the file and write first fibonacci numbers (0 and 1 ), then do fork and its child process read the last two numbers add them up and write the next into file and close the file and fork again this process continue like that with forking and child adding up numbers and writing calculated number into file, using fork inside the for not a good solution neither recursive call,is there any suggesstion for problem ??

Here is the link of the problem we are talking about multi-process part of the problem which is part 2

http://cse.yeditepe.edu.tr/~sbaydere/fall2010/cse331/files/assignments/F10A1.pdf

Answer 1

Assuming you're calculating them in a "simple" way (ie without using a cunning formula), I don't think it's a good candidate for parallel processing at all.

It's easy to come up with an O(n) solution, but each result depends on the previous one, so it's inherently tricky to parallelize. I can't see any benefit in your current parallel approach, as after each process has finished its own job and forked a child to get the next number, it's basically done... so you might as well do the work of the forked child in the existing process.

Answer 2

Fibonacci number calculation is a really strange idea to go multiprocess. Indeed, to calculate a number, you do need to know the previous two. Multiple processes cannot calculate other numbers but the next one, and only the next one. Multiple processes will all calculate the next Fibonacci number. Anyway, you'll double check.

Answer 3

You might want to look at this article:

http://cgi.cse.unsw.edu.au/~dons/blog/2007/11/29

There are more ideas here:

http://www.haskell.org/haskellwiki/The_Fibonacci_sequence

Answer 4

可能这不是解决你的问题，但这是计算给定范围的斐波纳契数的一种简单方法

int fibo(int n) { return (n<=2)?n:(fibo(n-1))+(fibo(n-2)); }