A Julia-based benchmarking project for comparing different matrix multiplication algorithms using high-performance computing (HPC) strategies. The tests are performed on the ClusterUY HPC system.
This project compares:
- A naive sequential algorithm.
- A multithreaded implementation using
Threads.@threads. - A distributed version using
MPI.jl. - Julia's built-in BLAS-optimized multiplication via
LinearAlgebra.
Matrix sizes are varied to study performance scaling. Benchmark results are stored in plain files and can be reused or analyzed externally.
Clone the repository and activate the environment:
julia --project
using Pkg
Pkg.instantiate()
## Automatización con Makefile
Este proyecto incluye un `Makefile` que permite ejecutar automáticamente benchmarks para varios métodos y tamaños de matrices.
Para correr todos los benchmarks definidos:
```bash
make