This README section will explain how to run the code in different modes.
The program to generate number arrays is in array_gen folder. The Makefile in it contains commands to create different types of arrays and save them in files. Comment off the unwanted ones and run make to automatically generate the files. Or simply run the ./generate_array command in the following format to generate a custom array of your choice.
./generate_array N distribution_type relative_output_file_name.bin
N could be any integer value. distribution_type should could be U for uniform distribution, N for normal distribution and E for exponential distribution.
To run the serial bucket sorting program on a generated set of numbers,
-
Navigate into
serialfolder and runmake. -
Run the
./bucketprogram in the folliwing format./bucket relative_input_file_path bucket_count -
If the code was run correctly, a file with the sorted set of numbers with the name
results.binwill be created in the same directory. -
Add the desired number of buckets in the place of
bucket_count.
To run the parallel bucket sorting program on a generated set of numbers,
-
Navigate into
parallelfolder and runmake. -
Run the
./bucketprogram in the folliwing format./bucket relative_input_file_path thread_count -
If the code was run correctly, a file with the sorted set of numbers with the name
results.binwill be created in the same directory. -
Add the desired number of buckets in the place of
thread_count. In the parallel implementation, thethread_countwill be taken as the number of buckets.
To validate if a generated file of numbers consists of accurately sorted numbers, navigate to check_accuracy folder and run make.
-
Run the
./check_accuracyprogram in the folliwing format./check_accuracy relative_path_of_result_file relative_path_of_original_file
The result will be displayed in the console.
The experiments are saved in a python notebook named experiments.ipynb in the experiments directory. It also contains the csv files which contains the data gathered for experimentation.
The directory can be reset (remove all generated files) by running make clean from the main directory path.
If required to clean files within the sub-directories, their versions of make clean can be used.