D2Q9 BGK Lattice Boltzmann code written in OpenCL and Modern Fortran.

An OpenCL demonstration code making use of the Focal OpenCL abstraction library for Fortran.

Getting started

Prerequisites

• A modern fortran compiler supporting the 2008 standard (tested with gfortran 9.1.0)
• An openCL development library (One of: Intel OpenCL SDK, NVIDIA CUDA Toolkit, AMD OpenCL SDK )

Download

This project makes use of git submodules to bundle a library dependency.

Therefore to clone the repository:

$> git clone git@github.com:LKedward/lbm2d_opencl.git
$> git submodule init
$> git submodule update

Build

To compile, navigate to the repository directory and run make:

$> make -j

Parallel build is fully supported so use of the -j flag is recommended. This will place executables in the ./bin/ directory.

To build the debug version:

$> make -j BUILD=debug

If the linker is unable to locate -lOpenCL you may have to specify the location of the OpenCL library (libOpenCL.so):

$> make -j OPENCL_DIR=/path/to/OpenCL/lib/

This should not be necessary on Linux platforms if the OpenCL development library is installed correctly

Run

To run the compiled program:

$> ./bin/lbmocl

Configuration file

When called with no arguments, the lbmocl program will run with a default configuration. To specify alternative configuration options, specify a configuration file as the first argument, e.g.:

$> ./bin/lbmocl lbm.cfg

The following optional parameters may be specified in the configuration file:

CASE: (string) which built-in test case to run. Must be one of: CAVITY, CYLINDER, STEP. Default: CAVITY.

NI: (integer) number of lattice points in vertical direction. Default: 400.

N_ITER: (integer) number of unsteady timestep iterations to perform. Default: 5000.

SAVE_FREQ: (integer) how often to calculate macroscopic variables and save to disk. Default: 100 (every 100 timesteps).

SAVE_FILE: (string) filename of tecplot output. Default: lbm.plt.

SAVE_TYPE: (string) tecplot save mode. Must be one of: BIN, TXT for binary (fast,small) and ascii (portable) modes respectively. Default: BIN (recommended).

CL_VENDOR: (string) which hardware vendor to select device. Default: NVIDIA.

The following options override parameters defined by the built-in test cases:

DX: (float) overrides test case lattice grid spacing.

DT: (float) overrides test case time step.

TAU: (float) overrides test case relaxation parameter.

About the code

Files and directories

• ./src/lbmocl.f90: main program code
• ./src/kernels.cl: OpenCL C kernel code implementing the lattice Boltzmann method
• ./src/UserInput.f90: module for processing user inputs from configuration file
• ./src/TecplotOutput.f90: module for writing solutions to tecplot files
• ./external/: directory containing code for external libraries
• ./makefile: makefile
• ./make.compiler: defines compiler flags, 'included' in makefile
• ./lbm.cfg: example configuration file containing parameter defaults

Acknowledgement

This work was funded by the MENtOR project, a UKVLN project supported by the Engineering and Physical Sciences Research Council (EPSRC) of the UK. Grant reference number EP/S010378/1