GPGPU Inverse Distance Weighting

A parallel GPU version of the IDW interpolation algorithm using matrix-vector multiplication [1].

custom_version -> matrix vector multiplication using a thread for each scalar product
cuBLAS_version -> cuBLAS matrix vector multiplication

Language: C, Framework: CUDA 8.0, cuBLAS v2

Authors

Mario Ruggieri

e-mail: mario.ruggieri@uniparthenope.it

Livia Marcellino

e-mail: livia.marcellino@uniparthenope.it

Installation and Usage

src/demo.cu is an example of usage on random data. It launches the CPU and GPU algorithms and shows for both the execution time.

Demo installation

cd src
make

In this way you are making demo binary file

Demo usage

1th argument: type of usage (1 = file, 2 = random data)
2th argument: (type 1) known 3D points dataset file / (type 2) number of known values
3th argument : (type 1) 2D locations file / (type 2) number of values to interpolate
4th argument : number of query locations per iteration
5th argument : number of CUDA block threads
6th argument: search radius

Examples:

./demo 1 dataset.txt grid.txt 5000 256 400
./demo 2 20000 300000 5000 256 400

CPU and GPU output are saved into the current directory.

Each line of the dataset file must have this layout: longitude;latitude;z;

For example:

13.70104167;40.84895833;-702.20;
13.70312500;40.84895833;-700.20;
...

Each line of the grid file (query locations) must have this layout: longitude;latitude;

For example:

13.70104167;40.55104167;
13.70154225;40.55104167;
...

Software uses single-precision floating-points.

Version

This is a beta version made for academic purposes.

Licensing

Please read LICENSE file.

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.

References

[1] L. Marcellino, R. Montella, S. Kosta, A. Galletti, D. Di Luccio, V. Santopietro, M. Ruggieri, M. Lapegna, L. D’Amore and G. Laccetti. Using GPGPU accelerated interpolation algorithms for marine bathymetry processing with on premises and cloud based computational resources. In 12th International Conference on Parallel Processing and Applied Mathematics, 2017, accepted

[2] Shepard D. , A two-dimensional interpolation function for irregularly-spaced data, Proceedings of the 1968 ACM National Conference. pp. 517–524