Module gpu_sparse_solver 

GPU sparse linear system solver (CPU mock implementation).

Provides CSR sparse matrix storage, SpMV, dot product, AXPY, conjugate gradient (CG), and preconditioned CG solvers that mirror GPU dispatch semantics while executing on the CPU.

Structs

GpuSparseSolverStats

Performance statistics for a GPU sparse solve.

SparseMatrixGpu

Compressed Sparse Row (CSR) matrix stored in GPU-friendly layout.

Functions

gpu_axpy

In-place AXPY: y ← a·x + y.

gpu_cg_solver

Conjugate gradient solver for symmetric positive-definite systems A·x = b.

gpu_dot

Dense dot product: ∑ aᵢ bᵢ.

gpu_jacobi_preconditioner

Build the Jacobi (diagonal) preconditioner for matrix mat.

gpu_pcg_solver

Preconditioned conjugate gradient solver.

gpu_spmv

Sparse matrix–vector product: y = A · x.

sparse_diagonal_matrix

Build a diagonal matrix from a dense vector.

sparse_identity

Build an n×n identity matrix in CSR format.