Fibonacci numbers with OpenMP tasks
Is there any benefit by using OpenMP to parallelize the Fibonacci number calculations? There are several examples online which calculate Fibonacci numbers using the task directive in OpenMP. For example at http://docs.oracle.com/cd/E19205-01/820-7883/girtd/index.html and here http://openmp.org/forum/viewtopic.php?f=3&t=1231 Some of these examples claim the performance is better with OpenMP. I don't understand this as calculating the Fibonacci series is, to my understanding, fundamentally non parallel (ignoring methods based on closed form solutions, e.g. from Binet's formula). Additionally,