DPCT1047#

Message#

The meaning of <parameter name> in the <API name> is different from the <API name>. You may need to check the migrated code.

Detailed Help#

In cuBLAS* and cuSolver* getrf API, the LU factorization is done as P*A=L*U; in Intel® oneAPI Math Kernel Library API, it is done as A=P*L*U. The result of the matrix P may be different.

Suggestions to Fix#

If the matrix P is only used in library API, ignore this warning. If P is used other ways, you may need to adjust the value of P.

For example, this original CUDA* code:

1 void foo(cublasHandle_t handle, float **a_array, float **b_array, int *p_array,
2          int *info_array) {
3   cublasSgetrfBatched(handle, 2, a_array, 2, p_array, info_array, 1);
4   cublasSgetrsBatched(handle, CUBLAS_OP_N, 2, 2, a_array, 2, p_array, b_array,
5                       2, info_array, 1);
6 }

results in the following migrated SYCL* code:

 1 void foo(dpct::queue_ptr handle, float **a_array, float **b_array, int *p_array,
 2          int *info_array) {
 3   /*
 4   DPCT1047:0: The meaning of p_array in the dpct::getrf_batch_wrapper is
 5   different from the cublasSgetrfBatched. You may need to check the migrated
 6   code.
 7   */
 8   dpct::getrf_batch_wrapper(*handle, 2, a_array, 2, p_array, info_array, 1);
 9   dpct::getrs_batch_wrapper(*handle, oneapi::mkl::transpose::nontrans, 2, 2,
10                             const_cast<float const **>(a_array), 2, p_array,
11                             b_array, 2, info_array, 1);
12 }

which is rewritten to:

 1 void foo(dpct::queue_ptr handle, float **a_array, float **b_array, int *p_array,
 2          int *info_array) {
 3   // cublas/cusolver API getrs need the output of cublas/cusolver getrf as input.
 4   // MKL/dpct helper API getrs need the output of MKL/dpct helper getrf as input.
 5   // In this case, matrix P is only used as a temporary data between library API
 6   // invocations, so the warning can be ignored.
 7   dpct::getrf_batch_wrapper(*handle, 2, a_array, 2, p_array, info_array, 1);
 8   dpct::getrs_batch_wrapper(*handle, oneapi::mkl::transpose::nontrans, 2, 2,
 9                             const_cast<float const **>(a_array), 2, p_array,
10                             b_array, 2, info_array, 1);
11 }