Multiplies a complex matrix by the complex unitary matrix Q determined by the hetrd (USM Version) function. This routine belongs to the mkl::lapack namespace.
cl::sycl::event unmtr(cl::sycl::queue &queue, mkl::side side, mkl::uplo uplo, mkl::transpose trans, std::int64_t m, std::int64_t n, T *a, std::int64_t lda, T *tau, T *c, std::int64_t ldc, T *scratchpad, std::int64_t scratchpad_size, const cl::sycl::vector_class<cl::sycl::event> &events = {});
unmtr (USM version) supports the following precisions and devices:
T | Devices supported |
---|---|
std::complex<float> | Host and CPU |
std::complex<double> | Host and CPU |
The routine multiplies a complex matrix C by Q or QH, where Q is the unitary matrix Q formed by hetrd (USM Version) when reducing a complex Hermitian matrix A to tridiagonal form: A = Q*T*QT. Use this routine after a call to hetrd (USM Version).
Depending on the parameters side and trans, the routine can form one of the matrix products Q*C, QT*C, C*Q, or C*QT (overwriting the result on C).
In the descriptions below, r denotes the order of Q:
r = m | if side = side::left |
r = n | if side = side::right |
Device queue where calculations will be performed.
Must be either side::left or side::right.
If side = side::left, Q or QT is applied to C from the left.
If side = side::right, Q or QT is applied to C from the right.
Must be either uplo::upper or uplo::lower. Uses the same uplo as supplied to hetrd (USM Version).
Must be either transpose::nontrans or transpose::trans.
If trans = transpose::nontrans, the routine multiplies C by Q.
If trans = transpose::trans, the routine multiplies C by QT.
The number of rows in the matrix C(m ≥ 0).
The number of columns in the matrix C(n ≥ 0).
The pointer to memory returned by hetrd (USM Version).
The leading dimension of a(max(1, r) ≤ lda).
The pointer to tau returned by hetrd (USM Version). The dimension of tau must be at least max(1, r-1).
The pointer to memory containing the matrix C. The second dimension of c must be at least max(1, n).
The leading dimension of c(max(1, n) ≤ ldc).
Pointer to scratchpad memory to be used by the routine for storing intermediate results.
Size of scratchpad memory as a number of floating point elements of type T. Size should not be less than the value returned by the unmtr_scratchpad_size function.
List of events to wait for before starting computation. Defaults to empty list.
Overwritten by the product Q*C, QT*C, C*Q, or C*QT (as specified by side and trans).
Exceptions
mkl::lapack::exception |
This exception is thrown when problems occur during calculations. You can obtain the info code of the problem using the info() method of the exception object: If info = -i, the i-th parameter had an illegal value. If info is equal to the value passed as scratchpad size, and detail() returns non-zero, then the passed scratchpad has an insufficient size, and the required size should not be less than the value returned by the detail() method of the exception object. |
Output event to wait on to ensure computation is complete.