IBM_TENO

IBM_TENO is a GPU-accelerated solver for simulating supersonic flows using a combination of the Immersed Boundary Method (IBM) and Targeted Essentially Non-Oscillatory (TENO) schemes. This solver is implemented in Fortran and is designed to handle complex geometries efficiently on CUDA-enabled GPUs.

Features

• GPU parallelized using NVIDIA CUDA Fortran
• Combines Immersed Boundary Method with high-order TENO schemes
• Capable of solving supersonic flow over complex geometries

Included Test Cases

The repository contains three .f90 source files, each corresponding to a specific benchmark case:

• Cylinder.f90: Supersonic flow around a single cylinder at Mach 3
• Multi_Cylinder.f90: Supersonic flow around multiple cylinders at Mach 2
• Schardin.f90: Simulation of Schardin's problem (shock interaction with a wedge)

Compilation

The solver is written in CUDA Fortran and can be compiled using the nvfortran compiler from the NVIDIA HPC SDK.

Compile Command:

nvfortran -cuda -O3 <source_file>.f90 -o <output_binary>.out

For example:

nvfortran -cuda -O3 Cylinder.f90 -o Cylinder.out

Contributing

We welcome any contributions and feedback that can help improve MHD_GDDC_GPU. If you would like to contribute to the tool, please contact the maintainers or open an Issue in the repository or a thread in Discussions. Pull Requests are encouraged, but please propose or discuss the changes in the associated Issue beforehand.

Licencing

Please refer to the licence file.