Parallel & distributed
sorting
One sorting problem, explored through four implementations — from a sequential baseline to execution across a cluster.
- Designed and implemented merge sort for variable-length records in C++.
- Benchmarked execution time, speedup, efficiency, and scalability across multiple cores and cluster nodes using SLURM job scripts.
- Documented the trade-offs between shared-memory and message-passing decomposition.
Four implementations
- 01Sequential
The reference baseline
- 02OpenMP
Shared-memory parallelism
- 03FastFlow
Structured parallel execution
- 04MPI + FastFlow
Hybrid distributed execution
Measured across cores & cluster nodes
