topic Hi, in Intel® oneAPI Math Kernel Library

fortran mkl csr sparse matrix storage, integer size of row/column index vectors for very large arrays

may_ka — Sat, 07 Oct 2017 11:40:23 GMT

Hi there,

I ran into trouble with the single precision integer size (I think it is 32bit) of the row index vector of mkl csr arrays. I create squared sparse arrays with +56,000,000 rows and columns and +3,000,000,000 elements. The column index vector is still ok because largest numbers are equal to the column dimension. However, the row index vector makes trouble with entries equal to the length of the column index vector which cannot be held by a single precision integer. I could use a larger integer kind number (double precision integer, 64 bit) but then I am in trouble with all mkl routines dealing with csr matrices (e.g. dcsrmm). That is because these routine have a fortran 77 interfaces and as outlined in the mkl manual, the expect index vectors of kind single precision. I can imagine that one could set the default integer to 64 bit when installing the mkl, but I am not sure whether that is possible. Any Ideas??

Thanks a lot

I found that "ilp64" is my

may_ka — Sat, 07 Oct 2017 11:43:53 GMT

I found that "ilp64" is my friend.

Seems to be solved.

Hi,

Xuhao_C_ — Tue, 28 Apr 2020 19:21:34 GMT

Hi,

I have the same problem. Basically the row indices are in a 64-bit array (A_idx_ptr) and the column indices are in a 32-bit array (A_nnz_idx).

mkl_scsrmm (&transa, &M , &N, &K, &alpha , matdescra, A_nonzeros, A_nnz_idx, A_idx_ptr, A_idx_ptr+1, B, &N, &beta , C, &N);

How did you call the above function?

Hi,

may_ka — Tue, 28 Apr 2020 22:19:34 GMT

Hi,

you just have to link you application against the version of mkl where the default integer is 64 bit. Checkout the mkl link adviser to get the link line correct.

Cheers

Hi, Currently we do not

MariaZh — Tue, 28 Apr 2020 23:03:20 GMT

Hi,
Currently we do not support mixing of different integer types for column and row indices in Sparse BLAS.
If you need long integers, you will have to convert column indices and use ILP64 interface of MKL.

Best regards,
Maria

Xuhao C. wrote:
Hi,
I have the same problem. Basically the row indices are in a 64-bit array (A_idx_ptr) and the column indices are in a 32-bit array (A_nnz_idx).
mkl_scsrmm (&transa, &M , &N, &K, &alpha , matdescra, A_nonzeros, A_nnz_idx, A_idx_ptr, A_idx_ptr+1, B, &N, &beta , C, &N);
How did you call the above function?

So you convert column indices

Xuhao_C_ — Wed, 29 Apr 2020 00:59:21 GMT

So you convert column indices into 64-bit before pass it to the function?

may.ka wrote:
Hi,
you just have to link you application against the version of mkl where the default integer is 64 bit. Checkout the mkl link adviser to get the link line correct.

Cheers

I see. So either both 32-bit

Xuhao_C_ — Wed, 29 Apr 2020 01:00:40 GMT

I see. So either both 32-bit or both 64-bit. Thanks.

Zhukova, Maria (Intel) wrote:
Hi,
Currently we do not support mixing of different integer types for column and row indices in Sparse BLAS.
If you need long integers, you will have to convert column indices and use ILP64 interface of MKL.
Best regards,
Maria
Quote:
Xuhao C. wrote:

Hi,
I have the same problem. Basically the row indices are in a 64-bit array (A_idx_ptr) and the column indices are in a 32-bit array (A_nnz_idx).
mkl_scsrmm (&transa, &M , &N, &K, &alpha , matdescra, A_nonzeros, A_nnz_idx, A_idx_ptr, A_idx_ptr+1, B, &N, &beta , C, &N);
How did you call the above function?

BTW, does csrmm support multi

Xuhao_C_ — Wed, 29 Apr 2020 21:44:48 GMT

BTW, does csrmm support multi-threading? Say if I export MKL_NUM_THREADS=32 on a 32-core CPU, will it get speedup compared to sequential setting?

Zhukova, Maria (Intel) wrote:
Hi,
Currently we do not support mixing of different integer types for column and row indices in Sparse BLAS.
If you need long integers, you will have to convert column indices and use ILP64 interface of MKL.
Best regards,
Maria
Quote:
Xuhao C. wrote:

Hi,
I have the same problem. Basically the row indices are in a 64-bit array (A_idx_ptr) and the column indices are in a 32-bit array (A_nnz_idx).
mkl_scsrmm (&transa, &M , &N, &K, &alpha , matdescra, A_nonzeros, A_nnz_idx, A_idx_ptr, A_idx_ptr+1, B, &N, &beta , C, &N);
How did you call the above function?

this routine should be

Gennady_F_Intel — Thu, 30 Apr 2020 12:27:50 GMT

this routine should be threaded, but the performance of csrmm functionality significantly depends on matrix size, sparsity, and portrait. Please note, this routine is deprecated. You may try to call Inspector-Executed counterpart (mkl_sparse_s_mm) of this routine from the latest version of mkl.