Skip to content

Implement ConjGrad, FwdTrans, and HelmSove operator for CUDA

Jacques Xing requested to merge CFD-Xing/redesign-prototypes:conjgrad-cuda into master

This MR implement the CUDA version of the ConjGrad, FwdTrans, and HelmSove operators.

Note: This preliminary CUDA implementation is not efficient. To be fixed in future MRs

Note: Robin boundary condition not yet implemented.

Edited by Jacques Xing

Merge request reports