import concurrent.futures as cf, multiprocessing ; from time import time                                                # Import python modukles for multiprocessing and timimg.
def multiply_on_a_CPUs_core(multiplications_on_this_CPU_core):                                                          # Define a function that multiplies floating point numbers on each CPU core.
    for n in range(multiplications_on_this_CPU_core):                                                                   # Look for the best number of multiplications that fit in this CPU's cache.
        x = n * 1.2345                                                                                                  # Woo hoo! Multiply!
    return                                                                                                              # Return from the multiplication function.
print("CPU cores used, total multiplications, multiplications per second")                                              # Describe the output columns.
for cpu_cores in range(1, 1 + multiprocessing.cpu_count()):                                                             # Test if more cores are faster.
    for exponent_of_batch_size_of_floating_point_numbers_to_multiply_on_a_CPU in range(30):                             # Consider increasing the range of exponents until multiplications per second peaks.
        total_multiplications = 2**exponent_of_batch_size_of_floating_point_numbers_to_multiply_on_a_CPU                # Convert the exponent to total mutiplications to bench mark.
        futures = []                                                                                                    # Create an empty python list for concurrent.futures.
        start_time = time()                                                                                             # Note when multiplications started.
        with cf.ProcessPoolExecutor() as executor:                                                                      # I don't know why this would be needed.
            for i in range(cpu_cores):                                                                                  # For each CPU core being used...
                futures.append(executor.submit(multiply_on_a_CPUs_core, total_multiplications // cpu_cores ))           # assign its share of of the mutiplications to it.
        print('%.i, %.i, %i' % (cpu_cores, total_multiplications, total_multiplications / ( time() - start_time ) ))    # Report the results.