Write this in c$++$

Final Assignment: Benchmarking Performance of the Sieve of Eratosthenes

Instead of a final exam, you will be given a final project, where you compare the performance of the

sieve of Eratosthenes with varying amounts of threads$/$processes and different sizes of workloads.

This assignment will consist of the following parts:

The multi$-$threaded sieve of the Eratosthenes program

MPI benchmarks

Answer these questions $($you can be as short or as detailed as you like$)$:

a$.$ Where do you think the "sweet spot" is for thread count? Alternatively, at what point

does the addition of more threads result in diminishing returns?

b$.$ Where do you think the "sweet spot" is for workload size? Alternatively, at what

maximum value $($max value to search for prime numbers$)$ does it become beneficial to

run the sieve in parallel?

Answer this additional question with a few sentences: Between this and the OpenMP$-$based

sieve from the midterm, what can you say are the advantages and disadvantages of both

OpenMP and MPI? You may consider the following:

a$.$ Which was easiest to program?

b$.$ Which one requires $($or benefits from$)$ specific hardware?

You will be benchmarking the sieve made using MPI. Start with the following table, consisting of

$12$ entries. You may expand the table to include different thread counts or maximum values.

A few notes:

If you have been following along using the video tutorials, you'll know that we cannot test a one$-$

process case since there must be at least one process for process $0$ to talk to$.$ Hence, the table has

$12$ entries instead of $16.$

If you were able to code a sieve that has process $0$ doing work alongside other processes,

then you can add in a one$-$process row.

You are welcome to present your findings as a graph rather than a table.

Try to run as many trials as possible $($at least $5$ trials, but more helps$)$ for each combination of

thread$/$process count and maximum number for primes and find the average execution time.