A newer version of this document is available. Customers should click here to go to the newest version.
Visible to Intel only — GUID: GUID-81A3B45A-BC50-49DB-B7BA-3FAF26538EB0
Visible to Intel only — GUID: GUID-81A3B45A-BC50-49DB-B7BA-3FAF26538EB0
oneapi::mkl::sparse::omatcopy
Performs an out-of-place copy/transpose of user data in a matrix handle into user data in a new handle without changing the sparse matrix format.
Description
The oneapi::mkl::sparse::omatcopy API copies user-provided sparse matrix arrays stored in a given sparse matrix handle into those provided in another sparse matrix handle. The routine enables users to perform a transpose operation and/or a change in array indexing in the copy. Only out-of-place copy/transpose is supported through this API.
API
Syntax
Using USM and SYCL buffers:
namespace oneapi::mkl::sparse { sycl::event omatcopy ( sycl::queue &queue, oneapi::mkl::transpose transpose_val, oneapi::mkl::sparse::matrix_handle_t spMat_in, oneapi::mkl::sparse::matrix_handle_t spMat_out, const std::vector<sycl::event> &dependencies = {}); }
Include Files
oneapi/mkl/spblas.hpp
Input Parameters
- queue
-
Specifies the SYCL command queue which will be used for SYCL kernels execution.
- transpose_val
-
Specifies operation op() on input matrix.
- oneapi::mkl::transpose::nontrans
-
Non-transpose, op(A) = A.
- oneapi::mkl::transpose::trans
-
Transpose, op(A) = AT.
- oneapi::mkl::transpose::conjtrans
-
Conjugate transpose, op(A) = AH.
- spMat_in
-
Handle to object containing input sparse matrix and other internal data. Created using one of the oneapi::mkl::sparse::set_<sparse_matrix_type>_data routines.
NOTE:The supported cases for <sparse_matrix_type> are csr and coo on both CPU and GPU devices. spMat_out must be created with the same <sparse_matrix_type>. - spMat_out
-
Handle to object containing output sparse matrix and other internal data. Created using one of the oneapi::mkl::sparse::set_<sparse_matrix_type>_data routines. Note that the sparse matrix dimensions/array lengths of this output handle must be same as that of the input handle in case of oneapi::mkl::transpose::nontrans operation. The number of rows and columns of this output matrix handle for oneapi::mkl::transpose::trans and oneapi::mkl::transpose::conjtrans operations must be the number of columns and rows of the input matrix handle, respectively. Also note that in case of a transpose operation, the number of non-zeros in the sparse matrix remains unchanged.
NOTE:The supported cases for <sparse_matrix_type> are csr and coo on both CPU and GPU devices. spMat_in must be created with the same <sparse_matrix_type>. - dependencies
-
A vector of type std::vector<sycl::event> & containing the list of events that the oneapi::mkl::sparse::omatcopy routine depends on.
Return Values
- sycl::event
-
SYCL event which can be waited upon or added as a dependency for the completion of the omatcopy routine.
// Examples of API arguments and usage sycl::event ev_opt = sparse::omatcopy(queue, trans_val, spMat_in, spMat_out); // Allowed use in case of USM sycl::event ev_opt = sparse::omatcopy(queue, trans_val, spMat_in, spMat_out, {}); // Allowed use in case of USM sycl::event ev_opt = sparse::omatcopy(queue, trans_val, spMat_in, spMat_out, dependencies); // Allowed use in case of USM sparse::omatcopy(queue, trans_val, spMat_in, spMat_out); // Recommended use in case of sycl::buffer sparse::omatcopy(queue, trans_val, spMat_in, spMat_out, {}); // Not recommended in case of sycl::buffer, but supported; sparse::omatcopy(queue, trans_val, spMat_in, spMat_out, dependencies); // Not recommended in case of sycl::buffer, but supported; sycl::event ev_opt = sparse::omatcopy(queue, trans_val, spMat_in, spMat_out); // Not recommended in case of sycl::buffer, but supported. Note that most sycl::buffer oneMKL APIs do not have ability to pass in dependencies