Backdrop through eigen-decomposition

Part of my model involves an eigen-decomposition, i.e. using torch.linalg.eigh, and I get the following error that I do not understand. Anyone has any ideas on what the error implies?

RuntimeError: cusolver error: CUSOLVER_STATUS_INTERNAL_ERROR, when calling cusolverDnXsyevd( handle, params, jobz, uplo, n, CUDA_R_32F, reinterpret_cast<void*>(A), lda, CUDA_R_32F, reinterpret_cast<void*>(W), CUDA_R_32F, reinterpret_cast<void*>(bufferOnDevice), workspaceInBytesOnDevice, reinterpret_cast<void*>(bufferOnHost), workspaceInBytesOnHost, info)

Could you post a minimal, executable code snippet which would reproduce the issue in the latest stable/nightly release, please?