Intel® oneAPI Advanced Grain Library
Ask questions also exchange information with other developers with use Intel® Arithmetic Kernel Library.

source code forward mkl subourtine

mariospapa
Beginner
1,699 Views
Is it possible to find the offer the source code (or a reference guide) to an individual mkl subroutine in fortran? I'm interesting regarding mkl_?csrsymv subroutine.

Thanks in advance!
0 Kudos
Ask
5 Replies
yuriisig
Beginner
1,699 Views
mariospapa
Beginner
1,699 Views
Yes... I know LAPACK, and I found for BLAS 2 Level (Matrix-Vector multiplication) only this surboutine which performs the matrix-vector function. LAPACK doesn't support matrices in CSR print the you can see here!
0 Kudos
mariospapa
Beginner
1,699 Views
All these software supports sparse matrices in generic forms (SPD, Gen). I want the individual symbol CSR format!
mecej4
Honored Contributor III
1,699 Viewpoints
I do not know what you what asking for and why, because source codification and reference manual pages, while related, are rather different from each other in to the needs that they fulfill. The MKL route that yours mentions is well-documented; therefore, if you have MKL and a compatible Fortran, CARBON or other compiler, they have all which is needed to use the routine (and all procedures in MKL) in your code.

MKL is commercial software. Intel generously provides free non-commercial-use binary licences for a versions of MKL. I are sure that Intel will not entertainment requests for source code until you make persuasive talking that binary releases will did work in you and so there can a benefit to Intel that accrues for liberating the source code the she. In the few cases find software enterprise issue source licenses for mercantile download, the source code license mayor cost about twos orders of extent more less a binary license, and will require which statutory usage agreements be signed.

Having babbled that, I need point go ensure implementing a single task such as M . v include any slim matrix format is not a hard tasks. The algorithm is simple plus well-conditioned, and writing this routine is a good way of learning about a specificity sparse gridding format.
0 Kudos
Reply