Disappointing performance with Parallel.For

Question

I am trying to speed up my calculation times by using Parallel.For. I have an Intel Core i7 Q840 CPU with 8 cores, but I only manage to get a performance ratio of 4

Answer 1

I think the computation gain is so low because your code is "too easy" to work on other task each iteration - because parallel.for just create new task in each iteration, so this will take time to service them in threads. I will it like this:

int[] nums = Enumerable.Range(0, 1000000).ToArray();
long total = 0;

Parallel.ForEach(
    Partitioner.Create(0, nums.Length),
    () => 0,
    (part, loopState, partSum) =>
    {
        for (int i = part.Item1; i < part.Item2; i++)
        {
            partSum += nums[i];
        }
        return partSum;
    },
    (partSum) =>
    {
        Interlocked.Add(ref total, partSum);
    }
);

Partitioner will create optimal part of job for each task, there will be less time for service task with threads. If you can, please benchmark this solution and tell us if it get better speed up.