/Parallel-DFT-CUDA-MPI

Parallel implementation of DFT on CUDA GPU and MPI architectures

Primary LanguageJupyter Notebook

Parallel DFT for the GPU and Software

This repository designs and evaluates a parallel DFT (Discrete Fourier Transform) implementation on CUDA GPUs and MPI (Message Passing Interface) architectures.

Implementation

CUDA MPI
CUDA Impl MPI Impl

Evaluation

Both CUDA and MPI implementations saw significant performance benefits over the sequential algorithm.

The full report can be found here.

Performance Graph showing better parallel performance