cuda-gen

This project is a small offshoot from Gen_PES_Pred(https://github.com/VaradRengaraj/Gen_PES_Pred). In Gen_PES_Pred, as part of the dataset generation process, a large file containing the coordinate frames of the atoms is parsed and for each frame, a coulomb matrix[1] is generated. This coulomb matrix is then sparsified and the resultant sparse coulomb matrix is divided into several dense smaller submatrices[2]. The eigenvalues for the submatrices are calculated and each of these eigenvalues vectors are stored in an HDF5 file. A single-threaded, serially running code executing this data generation process is time-consuming and hence a multithreaded Cuda-based approach is explored.

Kernels

Kernel Name	Description
coulomb	Generates coulomb matrix
submatrix	Sparse coulomb matrix is divided
	to smaller submatrices
jacobi-eigen	Calculates eigen values for a
	symmetric matrix

This is an ongoing project

References

[1] https://singroup.github.io/dscribe/0.3.x/tutorials/coulomb_matrix.html.
[2] Lass, M.; Mohr, S.; Wiebeler, H.; Kühne, T.D.; Plessl, C. "A Massively Parallel Algorithm for the Approximate Calculation of Inverse P-th Roots of Large Sparse Matrices". In Proceedings of the Platform for Advanced Scientific Computing Conference, pp. 1-11. 2018.