|
Uni20
Asynchronous tensor and runtime library.
|
Logical mdspan matrix layout before provider-specific BLAS lowering. More...
#include <mdspan_matrix.hpp>
Public Attributes | |
| Handle | data {} |
| blas_int | extent0 = 0 |
| blas_int | extent1 = 0 |
| bool | needs_conjugation = false |
| blas_int | nonunit_stride = 0 |
| Mdspan stride orthogonal to the unit-stride axis. | |
| int | unit_stride_axis = 0 |
Logical mdspan matrix layout before provider-specific BLAS lowering.
| Handle uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::data {} |
| blas_int uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::extent0 = 0 |
| blas_int uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::extent1 = 0 |
| bool uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::needs_conjugation = false |
| blas_int uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::nonunit_stride = 0 |
Mdspan stride orthogonal to the unit-stride axis.
| int uni20::linalg::blas::MdspanMatrixStage< Scalar, Handle >::unit_stride_axis = 0 |