In our CSDS 438 HPC final project we measure the effect of adding more processing resources to commonly implemented sorting algorithms.
STATUS: Working!
RUNNING MERGESORT
After Cloning:
To run MergeSort, simply cd to the Mergesort file cd ./mergeSort and submit the sorting.slurm job file to the HPC enviroment by running sbatch sorting.slurm. When the procedure is done, you should see an output file in this folder named something like <out.\w+>. This files contain the results.
CSV
The CSV file was made by hand using the outputs of the SLURM script described above. We do not currently have a scripted way to produce a CSV file from MergeSort output.
STATUS: Working!
RUNNING QUICKSORT
After Cloning:
To run QuickSort, simply cd to the Mergesort file cd ./quickSort and submit the qs-job.slurm job file to the HPC enviroment by running sbatch qs-job.slurm. When the procedure is done, you should see an output file in this folder named something like <out.\w+>. This files contain the results.
CSV
The CSV file was made by hand using the outputs of the SLURM script described above. We do not currently have a scripted way to produce a CSV file from QuickSort output.
STATUS: Working, nonparallizable
RUNNING HEAPSORT
After Cloning:
To run HeapSort, simply cd to the Mergesort file cd ./heapSort and compile and execute on a compute node in HPC.
STATUS: Working with slight modifications, not reasonably implementable in OpenMP
RUNNING SHELLSORT coming with next version of OMP. However, you can find a working serial program of shellsort and the skeleton of a parallisable one in the ./shellSort folder.
Please refer to the paper
Please email one of the authors in the title of paper