Intel® Math Kernel Library

PARDISO-Solving a large matrix containing many identical nonzero sparsity structure submatrices

Hi friends,

I am solving coupled problem analysis in FEM. I want to solve a matrix A.x = b, A could be

A = [A11, A12;  A21, A22] and B has a size of [A11, A12]

A could contain 3x3, 4x4...nxn submatrices. The submatrices have the same nonzero sparsity structure (IA and JA). At the moment the solution may be reordering IA and JA from submatrices to a large matrix A to input into Pardiso Solver. However it will increase the storage memory for IA and JA.

Wrong Schur complement


I'm using Pardiso to compute the Schur complement of a symmetric positive-semidefinite matrix with a kernel of dimension 6. I don't know if this is officially supported, but for similar rank-deficient matrices, I sometimes get the correct results. To see whether the Schur complement is right or not, I'm checking in MATLAB the eigenvalues of the result given by Pardiso after the numerical factorization. The first 6 eigenvalues should be 0. With MUMPS, I get the following (correct) output:

>> eig(triu(MUMPS) + tril(MUMPS') - diag(diag(MUMPS)))

ans =

link error: libmkl_core.a depends on Open MPI (via libmkl_blacs_openmpi_lp64.a)

While invoking 2015.2.164 Intel icpc compiler, I have encountered a link error while linking against libmkl_core.a:

/opt/intel/composer_xe_2015.2.164/mkl/lib/intel64/libmkl_core.a(cpardiso_blacs_lp64.o): In function 'mkl_pds_lp64_cpardiso_mpi_barrier':

__work/lnx32e/_cpardiso/kernel/mpi_wrapper/cpardiso_blacs_lp64_h.f:(.text+0x6): undefined referece to 'MKL_Barrier'

Installation fails --Jave Class Not Found -- on MacBook Pro

I am trying to install the Math Kernel Library under: 

Product Subscription Information
Download Latest Update
Release Posted

Academic Research Performance Libraries from Intel (OS X*)


On my MacBook Pro, when I try to start the install shell, I receive the following message:


In JPanelLicenseOptions

Inside JPanelRegistrationBegin

Calling initComponents

Calling initializePanel

Complex 1-D DFT not respecting DFTI_THREAD_LIMIT


I recently noticed that when using a threaded 1-dimensional DFT, a DFTI_COMPLEX domain DFT does not appear to respect the DFTI_THREAD_LIMIT and instead always uses the threading value set by mkl_set_num_threads().  Furthermore, it appears that while a REAL domain DFT does obey DFTI_THREAD_LIMIT, its behavior has changed between MKL v11.1 and 11.2.

Slow rectangular matrix transposition ?


I'm working with MKL on Gentoo. I have an "Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz" processor.

I'm trying to speed my inplace matrix transpositions and for that I thought that mkl_?imatcopy would be the solution. I have a very speedup on square matrix, but on rectangular matrix it is much worse than my naive "follow the cycles" implementation.

Here is the call:

mkl_dimatcopy('R', 'T', rows, cols, 1.0, matrix_ptr, rows, cols);

When I profiled the executable, most of the cycles were spent in

订阅 Intel® Math Kernel Library