Examples With Symmetric/Hermitian Eigenvalue Solvers

All examples are in 04_Symmetric_Eigenvalues/. Block and thread execution variants are provided.

• htev_batched_block.cu, htev_batched_thread.cu

These examples compute the eigenvalues and, optionally, the eigenvectors of batched symmetric/Hermitian tridiagonal matrices. Results are verified by checking \(A V = V \Lambda\), where \(V\) holds the eigenvectors and \(\Lambda\) is the diagonal matrix of eigenvalues.

• heev_batched_block.cu, heev_batched_thread.cu

These examples compute the eigenvalues and, optionally, the eigenvectors of batched symmetric/Hermitian matrices. Eigenvalues are compared with the cuSolver host API cusolverDnXsyevBatched and cusolverDn<t>syevjBatched; eigenvectors are verified via \(A V = V \Lambda\).

• hegst_batched_block.cu, hegst_batched_thread.cu

These examples reduce batched symmetric/Hermitian-definite generalized eigenproblems to standard form with HEGST. The result is verified against a cuBLAS reference for the chosen EigType and fill mode.

• hegv_batched_block.cu

These examples compute the eigenvalues and, optionally, the eigenvectors of batched symmetric/Hermitian-definite generalized eigenvalue problems with HEGV. Results are compared with cuSolverDn<t>sygvj or cuSolverDn<t>sygvd. Eigenvectors are verified using cuBLAS GEMM operations.