Skip to content
Snippets Groups Projects
Select Git revision
  • suffa/psm_optimization
  • fhennig/pystencils2.0-compat
  • master default protected
  • Sparse
  • WallLaw
  • improved_comm
  • release/1.3.7
  • release/1.3.6
  • release/1.3.5
  • release/1.3.4
  • release/1.3.3
  • release/1.3.2
  • release/1.3.1
  • release/1.3
  • release/1.2
  • release/1.1.1
  • release/1.1
  • release/1.0.1
  • release/1.0
  • release/0.4.4
  • release/0.4.3
  • release/0.4.2
  • release/0.4.1
  • release/0.4.0
  • release/0.3.4
  • release/0.3.3
26 results
Compare
Martin Bauer's avatar
Improvements for GPU code generation
Martin Bauer authored 
- turned on restrict keyword by default (makes large difference on GPUs)
- smarter block indexing: changing block size depending on domain size
  Example: previously there where (1,1,1) blocks when requested
  block size was (64, 1, 1) and domain size (1, 512, 512), now the
  block size is changed automatically to (1, 64, 1) in this case
- added __lauch_bounds__ to kernels to allow better optimizations from
  the CUDA compiler
9748ab43
History
Martin Bauer's avatar 9748ab43
History
Name Last commit Last update
doc
lbmpy
lbmpy_tests
COPYING.txt
MANIFEST.in
README.md
setup.py
README.md

lbmpy

Run fast fluid simulations based on the lattice Boltzmann method in Python.

alt text

Installation

export PIP_EXTRA_INDEX_URL=https://www.walberla.net/pip
pip install lbmpy[interactive]

Without [interactive] you get a minimal version with very little dependencies.

All options:

  • gpu: use this if nVidia GPU is available and CUDA is installed
  • alltrafos: pulls in additional dependencies for loop simplification e.g. libisl
  • interactive: installs dependencies to work in Jupyter including image I/O, plotting etc.

Options can be combined e.g.

pip install pystencils[interactive,gpu,doc]

Documentation

Read the docs here