Research-Grade GPU Kernel for Yee-Staggered Constrained Transport Hall-MHD: Advanced EMF Solver with MC Limiter and Full Pressure Tensor

This is a research-grade, highly optimized CUDA RawKernel (`ct_emf_kernel`) for computing the electromotive force (EMF) in 3D Hall-magnetohydrodynamic (Hall-MHD) simulations on a Yee-staggered grid using the Constrained Transport (CT) method. ### Key Features: - **Yee-Staggered Constrained Transport** implementation for maintaining ∇·B = 0 to machine precision - **Monotonized Central (MC) limiter** for robust velocity reconstruction at cell edges - **Hall term** fully incorporated via velocity correction using computed current density J - **Full anisotropic pressure tensor support** (includes off-diagonal components pe_xy, pe_xz, pe_yz) - **Adaptive numerical dissipation** based on local flow velocity - Efficient shared memory tiling with vectorized cooperative halo loading for high performance on modern GPUs - Designed for use with CuPy / CUDA Python workflows The kernel computes the three components of the electric field (Emfx, Emfy, Emfz) at edge centers, making it directly compatible with standard constrained transport updates for the magnetic field. This implementation is suitable for high-fidelity plasma simulations in astrophysics, space physics, and laboratory plasma research. It balances numerical stability, accuracy, and GPU performance. **Programming Language**: CUDA C++ (exposed via CuPy RawKernel) **Dependencies**: CuPy, NVIDIA GPU with compute capability ≥ 6.0