hegvd

Computes all eigenvalues and, optionally, eigenvectors of a real generalized symmetric definite eigenproblem using a divide and conquer method.

Description

hegvd supports the following precisions.

T
std::complex<float>
std::complex<double>

The routine computes all the eigenvalues, and optionally, the eigenvectors of a complex generalized Hermitian positive-definite eigenproblem, of the form

\(Ax = \lambda Bx, ABx = \lambda x\), or \(BAx =\lambda x\).

Here \(A\) and \(B\) are assumed to be Hermitian and \(B\) is also positive definite.

It uses a divide and conquer algorithm.

hegvd (Buffer Version)

Syntax

namespace oneapi::mkl::lapack {
  void hegvd(sycl::queue &queue, std::int64_t itype, oneapi::mkl::job jobz, oneapi::mkl::uplo upper_lower, std::int64_t n, sycl::buffer<T,1> &a, std::int64_t lda, sycl::buffer<T,1> &b, std::int64_t ldb, sycl::buffer<realT,1> &w, sycl::buffer<T,1> &scratchpad, std::int64_t scratchpad_size)
}

Input Parameters

queue

The queue where the routine should be executed.

itype

Must be 1 or 2 or 3. Specifies the problem type to be solved:

if \(\text{itype} = 1\), the problem type is \(Ax = \lambda Bx;\)

if \(\text{itype} = 2\), the problem type is \(ABx = \lambda x;\)

if \(\text{itype} = 3\), the problem type is \(BAx = \lambda x\).

jobz

Must be job::novec or job::vec.

If jobz = job::novec, then only eigenvalues are computed.

If jobz = job::vec, then eigenvalues and eigenvectors are computed.

upper_lower

Must be uplo::upper or uplo::lower.

If upper_lower = uplo::upper, a and b store the upper triangular part of \(A\) and \(B\).

If upper_lower = uplo::lower, a and b stores the lower triangular part of \(A\) and \(B\).

n

The order of the matrices \(A\) and \(B\) (\(0 \le n\)).

a

Buffer, size a(lda,*) contains the upper or lower triangle of the Hermitian matrix \(A\), as specified by upper_lower.

The second dimension of a must be at least \(\max(1, n)\).

lda

The leading dimension of a; at least \(\max(1,n)\).

b

Buffer, size b(ldb,*) contains the upper or lower triangle of the Hermitian matrix \(B\), as specified by upper_lower.

The second dimension of b must be at least \(\max(1, n)\).

ldb

The leading dimension of b; at least \(\max(1,n)\).

scratchpad_size

Size of scratchpad memory as a number of floating point elements of type T. Size should not be less than the value returned by hegvd_scratchpad_size function.

Output Parameters

a

On exit, if jobz = job::vec, then if \(\text{info} = 0\), a contains the matrix \(Z\) of eigenvectors. The eigenvectors are normalized as follows:

if \(\text{itype} = 1\) or \(\text{itype} = 2\), \(Z^{H}BZ = I\);

if \(\text{itype} = 3\), \(Z^{H}B^{-1}Z = I\);

If jobz = job::novec, then on exit the upper triangle (if upper_lower = uplo::upper) or the lower triangle (if upper_lower = uplo::lower) of \(A\), including the diagonal, is destroyed.

b

On exit, if \(\text{info} \le n\), the part of b containing the matrix is overwritten by the triangular factor \(U\) or \(L\) from the Cholesky factorization \(B = U^{H}U\)or \(B = LL^{H}\).

w

Buffer, size at least \(n\). If \(\text{info} = 0\), contains the eigenvalues of the matrix \(A\) in ascending order.

scratchpad

Buffer holding scratchpad memory to be used by routine for storing intermediate results.

hegvd (USM Version)

Syntax

namespace oneapi::mkl::lapack {
  sycl::event hegvd(sycl::queue &queue, std::int64_t itype, oneapi::mkl::job jobz, oneapi::mkl::uplo upper_lower, std::int64_t n, T *a, std::int64_t lda, T *b, std::int64_t ldb, RealT *w, T *scratchpad, std::int64_t scratchpad_size, const std::vector<sycl::event> &events = {})
}

Input Parameters

queue

The queue where the routine should be executed.

itype

Must be 1 or 2 or 3. Specifies the problem type to be solved:

if \(\text{itype} = 1\), the problem type is \(Ax = \lambda Bx;\)

if \(\text{itype} = 2\), the problem type is \(ABx = \lambda x;\)

if \(\text{itype} = 3\), the problem type is \(BAx = \lambda x\).

jobz

Must be job::novec or job::vec.

If jobz = job::novec, then only eigenvalues are computed.

If jobz = job::vec, then eigenvalues and eigenvectors are computed.

upper_lower

Must be uplo::upper or uplo::lower.

If upper_lower = uplo::upper, a and b store the upper triangular part of \(A\) and \(B\).

If upper_lower = uplo::lower, a and b stores the lower triangular part of \(A\) and \(B\).

n

The order of the matrices \(A\) and \(B\) (\(0 \le n\)).

a

Pointer to array of size a(lda,*) containing the upper or lower triangle of the Hermitian matrix \(A\), as specified by upper_lower. The second dimension of a must be at least \(\max(1, n)\).

lda

The leading dimension of a; at least \(\max(1,n)\).

b

Pointer to array of size b(ldb,*) containing the upper or lower triangle of the Hermitian matrix \(B\), as specified by upper_lower. The second dimension of b must be at least \(\max(1, n)\).

ldb

The leading dimension of b; at least \(\max(1,n)\).

scratchpad_size

Size of scratchpad memory as a number of floating point elements of type T. Size should not be less than the value returned by hegvd_scratchpad_size function.

events

List of events to wait for before starting computation. Defaults to empty list.

Output Parameters

a

On exit, if jobz = job::vec, then if \(\text{info} = 0\), a contains the matrix \(Z\) of eigenvectors. The eigenvectors are normalized as follows:

if \(\text{itype} = 1`\) or \(\text{itype} = 2\), \(Z^{H}BZ = I\);

if \(\text{itype} = 3\), \(Z^{H} B^{-1} Z = I\);

If jobz = job::novec, then on exit the upper triangle (if upper_lower = uplo::upper) or the lower triangle (if upper_lower = uplo::lower) of \(A\), including the diagonal, is destroyed.

b

On exit, if \(\text{info} \le n\), the part of b containing the matrix is overwritten by the triangular factor \(U\) or \(L\) from the Cholesky factorization \(B = U^{H}U\)or \(B\) = \(LL^{H}\).

w

Pointer to array of size at least n. If \(\text{info} = 0\), contains the eigenvalues of the matrix \(A\) in ascending order.

scratchpad

Pointer to scratchpad memory to be used by routine for storing intermediate results.

Return Values

Output event to wait on to ensure computation is complete.

Parent topic: LAPACK Singular Value and Eigenvalue Problem Routines