Back
Developement of a parallel adaptive solver
A block-based adaptive mesh refinement (AMR) library with multigrid poisson solver for the in-house CFD solver was developed with an aim of developing a massively parallelable physics-based adaptive two-phase flow solver. The library is written in C++ and extensively uses STL. Parallelization is done using MPI and was tested on the supercomputer (SahasraT - Cray XC40).
This project was carried out under the guidance of Prof. Gaurav Tomar.
Speedup plot
Gradient-based refinement with Block nesting (adjacent block refinement ratio = 2)