#Exercise: WorkingWithMultipleProcesses-8 Run the Fibonacci code on your machine. Do you get comparable timings? If your machine has multiple cores and/or processors, do you see improvements in the timing as we increase the application's concurrency?
See the fibonacci.exs file for the full modules.
So, I see no performance increase whatsoever:
iex> Runner.run
[{37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}]
# time (s)
1 5.90
2 5.85
3 5.86
4 5.91
5 5.97
6 5.87
7 5.86
8 5.90
9 5.86
10 5.84
:ok
I'm on a multi-core machine; however, I'm running this in a VM that is only allocated 1 processor. Maybe when I'm at home I'll try running this again. I would expect the results to be more in line with those in the book.
I'm a little confused why the results in the book show: {37, 39088169}; fib(37) should equal 24,157,817. Oh well.
So, the previous run was on my work computer but I tried the same thing again at home and saw interesting results:
iex> Runner.run
[{37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}, {37, 24157817}]
# time (s)
1 7.90
2 4.11
3 2.76
4 2.72
5 2.90
6 1.93
7 1.91
8 1.95
9 1.93
10 1.91
:ok
The single processor run was actually slower; however, adding processes significantly increased the performance. My home PC is running an i7 2600k, which has 4 cores and 8 threads, so it makes sense that there were performance gains. I'm unsure of why it levels off between 3 and 6 and then at 7 instead of 8 but regardless, it's pretty cool to see the power of concurrency. From 7.9 down to ~1.9 seconds is a pretty awesome increase.