Use mkl_sparse_convert_csr to transpose doesn't change rows & cols

yan__zixi — Tue, 15 Jan 2019 15:10:18 GMT

Hello, I am using mkl_sparse_convert_csr to perform transpose function by calling with SPARSE_OPERATION_TRANSPOSE option.

However, the values of the result matrix is transposed correctly with the row & col left unchanged.

Here is the code.

#include <bits/stdc++.h>
#include <mkl.h>
#include <omp.h>
using namespace std;

namespace {
// Output MKL sparse matrix in CSR format.
void Print(const sparse_matrix_t &mat) {
sparse_index_base_t index_base;
MKL_INT rows, cols;
MKL_INT *rows_start, *rows_end, *col_idx;
float *values;
assert(mkl_sparse_s_export_csr(mat, &index_base, &rows, &cols, &rows_start,
&rows_end, &col_idx,
&values) == SPARSE_STATUS_SUCCESS);
cout << "IndexBase=" << (index_base == SPARSE_INDEX_BASE_ZERO ? 0 : 1)
<< endl;
cout << "Rows=" << rows << " Cols=" << cols << endl;
for (int i = 0; i < rows; ++i) {
cout << "RowsKey=" << i << ":";
for (int j = rows_start; j < rows_end; ++j) {
cout << "(" << col_idx << "," << values << ")";
}
cout << endl;
}
}

} // namespace

int main() {
// Auto log flush.
cout.setf(std::ios::unitbuf);
mkl_set_num_threads(1);

const MKL_INT n = 4, m = 5;

// Create matrix a.
vector<float> values = {1, 2, 2, 1};
vector<MKL_INT> rows_start = {0, 0, 2, 3};
vector<MKL_INT> rows_end = {0, 2, 3, 4};
vector<MKL_INT> col_idx = {0, 2, 2, 3};

sparse_matrix_t mat_a;
assert(mkl_sparse_s_create_csr(&mat_a, SPARSE_INDEX_BASE_ZERO, n, m,
rows_start.data(), rows_end.data(),
col_idx.data(),
values.data()) == SPARSE_STATUS_SUCCESS);
cout << "Matrix A:" << endl;
Print(mat_a);

sparse_matrix_t mat_b;
assert(mkl_sparse_convert_csr(mat_a, SPARSE_OPERATION_TRANSPOSE, &mat_b) ==
SPARSE_STATUS_SUCCESS);
cout << "Matrix B:" << endl;
Print(mat_b);

mkl_sparse_destroy(mat_b);
mkl_sparse_destroy(mat_a);
return 0;
}

The result is here:

Matrix A:
IndexBase=0
Rows=4 Cols=5
RowsKey=0:
RowsKey=1:(0,1)(2,2)
RowsKey=2:(2,2)
RowsKey=3:(3,1)
Matrix B:
IndexBase=0
Rows=4 Cols=5
RowsKey=0:(1,1)
RowsKey=1:
RowsKey=2:(1,2)(2,2)
RowsKey=3:(3,1)

As you can see, the rows=4 & cols=5 in the matrix B.

Any idea what the problem is? Is there any useful information?

Thank you very much in advance.

Zixi

I wonder the behaviour of mkl

yan__zixi — Wed, 16 Jan 2019 12:29:00 GMT

I wonder the behaviour of mkl_sparse_convert_csr.

Is the result meet expectation?

I don't see anything about it in the document.

Is it a bug in mkl_sparse

yan__zixi — Fri, 15 Feb 2019 08:22:15 GMT

Is it a bug in mkl_sparse_convert_csr? Any progress?

Yes, this is the bug with

Gennady_F_Intel — Mon, 18 Feb 2019 03:32:56 GMT

Yes, this is the bug with converter in the case of transpose and conjugate cases. The issue is escalated. We will update the status of this issue asap.

Hello,

yan__zixi — Mon, 18 Feb 2019 03:42:19 GMT

Hello,

Thanks for your confirmation.

Now I use mkl_sparse_add to perform the transpose.

Thanks,

Zixi

The fix of the issue

Gennady_F_Intel — Mon, 16 Sep 2019 04:51:47 GMT

The fix of the issue available in MKL 2019 u5 which we released the last week. You may check this update and let us know how this update will work on your side.

topic I wonder the behaviour of mkl in Intel® oneAPI Math Kernel Library

Use mkl_sparse_convert_csr to transpose doesn't change rows & cols

I wonder the behaviour of mkl

Is it a bug in mkl_sparse

Yes, this is the bug with

Hello,

The fix of the issue