https://github.com/rkv0id/boltzmanumba

GPU-Parallelization of a sequential Lattice Boltzmann gist on CUDA-capable devices using Numba.
https://github.com/rkv0id/boltzmanumba

cuda lbm numba

Last synced: 28 days ago
JSON representation

GPU-Parallelization of a sequential Lattice Boltzmann gist on CUDA-capable devices using Numba.

• Host: GitHub
• URL: https://github.com/rkv0id/boltzmanumba
• Owner: rkv0id
• Created: 2021-03-01T10:35:17.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2021-03-05T18:27:09.000Z (almost 4 years ago)
• Last Synced: 2024-11-09T06:38:19.532Z (3 months ago)
• Topics: cuda, lbm, numba
• Language: Python
• Homepage:
• Size: 1.07 MB
• Stars: 3
• Watchers: 2
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# boltzmanumba



Parallelization of a sequential Lattice Boltzmann code picked up from a random GitHub gist.

LBM is a simulation technique for complex fluid systems. It is known to be so performance greedy while also being an embarassingly parallel algorithm. This repo is a basic attempt of distributing the LBM algorithm on a CUDA-able GPU using the [Numba API](http://numba.pydata.org).

The `lattice_gpu_naive.py` file contains a naive parallelization of the code while the `lattice_gpu_opt.py` file contains a *memory-usage-optimized* code for the same algorithm.

The notebook provided within this repository contains a benchmarking of the GPU/CPU acceleration.