rocsolver_sgehrd_batched Interface Reference

rocsolver_sgehrd_batched Interface Reference#

hipFORT API Reference: hipfort_rocsolver::rocsolver_sgehrd_batched Interface Reference
hipfort_rocsolver::rocsolver_sgehrd_batched Interface Reference

The GEHRD_BATCHED functions compute the upper Hessenberg form of a batch of general square matrices. More...

Public Member Functions

integer(c_int) function rocsolver_sgehrd_batched_ (handle, n, ilo, ihi, a, lda, tau, stridep, batch_count)
 
integer(kind(rocblas_status_success)) function rocsolver_sgehrd_batched_rank_0 (handle, n, ilo, ihi, a, lda, tau, stridep, batch_count)
 
integer(kind(rocblas_status_success)) function rocsolver_sgehrd_batched_rank_1 (handle, n, ilo, ihi, a, lda, tau, stridep, batch_count)
 
integer(kind(rocblas_status_success)) function rocsolver_sgehrd_batched_strided_rank (handle, n, ilo, ihi, a, lda, tau, stridep, batch_count)
 

Detailed Description

The GEHRD_BATCHED functions compute the upper Hessenberg form of a batch of general square matrices.

(This is the blocked version of the algorithm.)

For each instance in the batch, the upper Hessenberg form is given by:

\[ H_l = Q_l^H A_l Q_l \]

where \(H_l\) is an upper Hessenberg matrix, and \(Q\) is an n -by-n orthogonal/unitary matrix represented as the product of \((ihi-ilo)\) Householder matrices

\[ Q_l = H_l(ilo)H_l(ilo+1)\cdots H_l(ihi-1) \]

Each Householder matrix \(H(i)\) is given by

\[ H_l^{}(i) = I - \text{tau}_l^{}[i] \cdot v_{l_i}^{} v_{l_i}^H \]

where the first i elements of the Householder vector \(v_{l_i}\) are zero, and \(v_{l_i}[i+1] = 1\).

Parameters
[in]handle- rocblas_handle.
[in]n- rocblas_int. n >= 0. The number of rows and columns of all the matrices A_l in the batch.
[in]ilo- rocblas_int. 1 <= ilo <= ihi. The starting 1-based index of the row and column to be reduced.
[in]ihi- rocblas_int. ilo <= ihi <= n. The ending 1-based index of the row and column to be reduced.
[in,out]A- Array of pointers to type. Each pointer points to an array on the GPU of dimension lda*n. On entry, the n-by-n matrices A_l to be reduced. On exit, the elements on and above the first subdiagonal contain the upper Hessenberg form H_l, and the elements below the first subdiagonal are the last ihi - i - 1 elements of Householder vector v_(l_i).
[in]lda- rocblas_int. lda >= n. Specifies the leading dimension of matrices A_l.
[out]tau- pointer to type. Array on the GPU (the size depends on the value of strideP). Contains the vectors tau_l of corresponding Householder scalars.
[in]strideP- rocblas_stride. Stride from the start of one vector tau_l to the next one tau_(l+1). There is no restriction for the value of strideP. Normal usage is strideP >= n-1.
[in]batch_count- rocblas_int. batch_count >= 0. Number of matrices in the batch.

Member Function/Subroutine Documentation

◆ rocsolver_sgehrd_batched_()

integer(c_int) function hipfort_rocsolver::rocsolver_sgehrd_batched::rocsolver_sgehrd_batched_ ( type(c_ptr), value  handle,
integer(c_int), value  n,
integer(c_int), value  ilo,
integer(c_int), value  ihi,
type(c_ptr), value  a,
integer(c_int), value  lda,
type(c_ptr), value  tau,
integer(c_int64_t), value  stridep,
integer(c_int), value  batch_count 
)

◆ rocsolver_sgehrd_batched_rank_0()

integer(kind(rocblas_status_success)) function hipfort_rocsolver::rocsolver_sgehrd_batched::rocsolver_sgehrd_batched_rank_0 ( type(c_ptr), value  handle,
integer(c_int)  n,
integer(c_int)  ilo,
integer(c_int)  ihi,
type(c_ptr), value  a,
integer(c_int)  lda,
real(c_float), target  tau,
integer(c_int64_t)  stridep,
integer(c_int)  batch_count 
)

◆ rocsolver_sgehrd_batched_rank_1()

integer(kind(rocblas_status_success)) function hipfort_rocsolver::rocsolver_sgehrd_batched::rocsolver_sgehrd_batched_rank_1 ( type(c_ptr), value  handle,
integer(c_int)  n,
integer(c_int)  ilo,
integer(c_int)  ihi,
type(c_ptr), value  a,
integer(c_int)  lda,
real(c_float), dimension(:), target  tau,
integer(c_int64_t)  stridep,
integer(c_int)  batch_count 
)

◆ rocsolver_sgehrd_batched_strided_rank()

integer(kind(rocblas_status_success)) function hipfort_rocsolver::rocsolver_sgehrd_batched::rocsolver_sgehrd_batched_strided_rank ( type(c_ptr), value  handle,
integer(c_int)  n,
integer(c_int)  ilo,
integer(c_int)  ihi,
type(c_ptr), value  a,
integer(c_int)  lda,
real(c_float), dimension(:,:), target  tau,
integer(c_int64_t)  stridep,
integer(c_int)  batch_count 
)

The documentation for this interface was generated from the following file: