Intel® Math Kernel Library

Welcome to the Intel MKL Forum

The Mission of the Intel Math Kernel Library Forum is to provide users and interested developers an open, non-judgmental place to discuss topics or issues relating to the Intel Math Kernel Library product. Because of the breadth of the product-BLAS, LAPACK, DFTs, ScaLAPACK, Vector Math Library, Vector Statistical Library, etc.-we anticipate that the forum will draw a wide and diverse audience. We certainly hope so.

Looking for Sparse Sover help


I am trying to test out the sparse solver and learn how to code with MKL. So I inputed a 5x5 sparse matrix and a x vector

A = 1 -1 -3 0 0, -2 5 0 0 0, 0 0 4 6 4, -4 0 2 7 0, 0 -8 0 0 5

b = 1.2, 1.2, 1.2, 1.2, 1.2

Since it is a small matrix I used mathmatica to try to confirm the output of a system of linear equations Ay = b. I came up with:

0.107447, 0.282979, -0.458511, 0.36383, 0.212766

using mkl_cspblas_dcootrsv i get

3.13371, 0.624, 0.214286, 0.171429, -0.24

I used the following parameters for MKL

uplo = l

static and dynamic linking

hi group,

I am a PhD student, and in my project I have to modify some subroutines of codes that were written in fortran. I have a basic knowledge of fortran as I do not have a background of compouter science.

I have two different codes written in fortran. Both have scripts for compiling and linking provided by the suppliers.

Linking Error using DSS

Dear Intel Team!

I have the following problem:

I am building a static library which includes a function that encapsulates a routine which uses the DSS exactly the way as shown in the example in the Intel Math Kernel Library Reference Manual.
Building the library is ok, no errors occur.

But when I want to use the build library from another program following linking error occurs:

DPPTRI is not parallelized

I had used DPOTRI to inverse a positive definite symmertric matrix. Recently I had to reduce the memory usage, so I started using packed storage for this matrix and DPPTRI to invert it. When I used DPOTRI, it was automatically parallelized, however, DPPTRI was not parallelized. Is it a bug of MKL? Or DPPTRI does not support parallelization?

P.S. I'm using MKL 10.0.2 and ICC 10.1.015 on a Core 2 Quad machine (em64t mode, kernel 2.6.18). OMP_NUM_THREADS is set to 4.

MKL 10 fails the standard LAPACK tests

When running the LAPACK tests that come with the reference LAPACK library from netlib with MKL 10 ( and, the tests produce a segmentation fault.

The tests check the error exits, so they pass parameters to LAPACK routines that are not valid to see if the routine will identify an error. This seems to be causing the segmentation faults in the iterative refinement (e,g, DGERFS) and condition estimation (e.g. DGECON) LAPACK routines.

A segmentation fault also occurs if the routines like DGECON are called with N=0, which is a valid imput parameter.



I'm trying to build using the MKL version above, and am having issues getting this error:

MKL FATAL ERROR: SO k/kernel/source/xormrq_omp.fxx;mkl_lapack_zunmrq;323;333;; not convinient for this processor.

when I run on a Intel Xeon system, having dynamically linked MKL. I then compiled up the pdepoissonc example, using "make soem64t", and I get the same error. The example runs fine on an opteron system.

If I compile with "make libem64t" it works on both opteron and Xeon processors.

Suscribirse a Intel® Math Kernel Library