Bibliothèque Intel® Math Kernel Library

memory leakage problem

Dear all,

I have a project with multiple times using MKL functions. After tens of times calling my program, the memory usage is larger and large (from initial 100 MB to latter more than 10 GB). So I use the Intel Inspector to check the location of memory leak. However, the Inspector can't find the location from my code but it is in MKL's functions. The original post is here:

https://software.intel.com/en-us/forums/intel-inspector-xe/topic/606912#...

 

11.3 Update 1 integer divide by zero when using MKL_DIRECT

I am using the "direct" calls set by MKL_DIRECT and am moving from MKL 11.2 to MKL 11.3 Update 1

I am finding my code is throwing a "integer divide  by zero" exception at this call when using 11.3 Update 1. The "calling" code is well tested and worked fine in MKL 11.2.

If I "turn off" MKL_DIRECT, the code again works as expected.

 

	else if (!AisN && BisN)
		mkl_dc_gemm_xx_mnk_pst(m, n, k, alpha, A, lda, B, ldb, beta, C, ldc, MKL_DC_MOV, MKL_DC_MOV, MKL_DC_MT, MKL_DC_MN);

 

Feast RCI with real factorizations

Hello,

I would like to try to use FEAST to solve the generalized eigenproblem (A-sB)x = 0, where A is symmetric, positive-definite, and B is symmetric (actually, it's a diagonal matrix), positive semi-definite.

I'm all setup to use my own storage for A and B, and my own tools to factorize the shifted matrix A-sB. However - and this may be a huge issue - my tools only allow me to factorize a real matrix - not a complex one.

Getting different output when calling the same FFT functions of MKL from C++ and Fortran

I got different results when calling same functions of MKL from intel C++ compiler and IVF separately. I'm in windows 10 64bit, and VS 2013. My C++ code is as follows:

#include <complex>
#include <iostream>
#include "mkl_dfti.h"
int main()
{
 using namespace std;
 complex <float>x[32];
 int  i;
 DFTI_DESCRIPTOR_HANDLE my_desc1_handle;
 MKL_LONG status;

 for (i = 0; i < 32; i++)
 {
  x[i] = ((i+1)*1.0, (i+1)*1.0);
 }

Nonsymmetric eigenvalue problem

Hello,

I'm trying to compute the right eigenvectors of a Hessenberg matrix. It seems I get the correct results when using the driver routine ?geev, but I can't obtain the same results with the computational routines ?hseqr and ?hsein (which should in the end be faster because there is no need to reduce A into Hessenberg form). Do you see what is wrong in the file dhsein.f ?

$ ifort dgeev.f  -lmkl_core -lmkl_intel_thread -liomp5 -lmkl_intel_lp64 && ./a.out < dhsein.d

Is there something wrong in DataFitting interpolate with user defined extrapolate call back routine?

Hi all,

I meet a strange problem in DataFitting Module while call df?InterpolateEx1D function. If I present the call back function for extrapolation , the result was wrong. But if set them to NULL, result is right.

My problem is:

x[] = {1, 1.5, 2.5, 3.5, 4.5, 5}, and y[] = {100, 120, 150, 170, 200, 200}, Interpolate method is stepwise constant interpolant. And I need found y value at x=4.

The following is the sample code. I hope that was something wrong in my code. Can you help me to explain it? thanks.

Question on BLAS const arguments

Hi, 

I have a couple of general questions concerning multiple const arguments in blas and blas extension calls (C++ user).

1) When one has 2 const data (i.e. pointer) arguments, is there any harm in using the same data ?

For example, in dgemm, can the A and B matrices be the same, obtaining so, A^2, A*A^t and A^t*A ?

Is it a case by case answer? If so, how can one know?

2) In the mkl_?imatcopy which is an in-place operation why does one need 2 leading dimension arguments? 

S’abonner à Bibliothèque Intel® Math Kernel Library