Intel® Math Kernel Library

Cluster FFT - undefined reference

Hi, this is driving me crazy:
Assembled the simplest application just to test FFT on cluster (Linux SUSE 10.0). Could not compile it, always undefined reference just to specific Cluster FFT functions. Regular (non-cluster) FFT works just fine, but I wanted to try running it in parallel on our new cluster,a nd besides this is just a testbed for redesigning some larger project with mkl. Please help.

(Intel compiler 10.1.08, mkl: tried both 10.0.08 and, the same behaviour)
Bellow: Source code, command line, compiler/linker error.

LAPACK LOGICAL arguments on 32- or 64-bit platforms

Please pardon me if this is elementary...

I'm trying to figure out what C/C++ type to use for LAPACK arguments that are billed as "LOGICAL". These are mostly bwork arrays, but also include selector arguments. I need to know this for building both 32-bit and 64-bit versions of our application.

The old f2c definition is as a long, but that isn't consistent on 64-bit platforms. Apple defines these arguments as long for 32-bit platforms and int on 64-bit.

The mkl_lapack.h header file has void * for those arguments, so it provides no guidance.

MKL Kernel calling wrong dll - CPU Dispatch issue

I have been working on a project using MKL, statically linked (except for openmp). I didn't make any code changes, but all of a sudden calling LAPACK function dgesv would result in a quasi-infinite loop. The problem only appears once the matrices involved reach a certain size. I did not try and determine the exact dimensions for which the factorization fails.

To solve the problem, I switched all the libraries to dynamically linked. I included the *p4n.dll files in the project build folder and that seemed to remove the limit on the size of the matrices.

Intel MKL Pardiso, LP64, and ILP64 Interfaces

The Intel MKL user guide and the Intel MKL-supplied Fortran 90/95 interface to Pardiso (mkl_pardiso.f90) appear to be in contradiction.

On page 3-5 of MKL User Guide it reads:

"You may alternatively use other 64-bit types for the integer parameters that must be
64-bit in ILP64. For example, with Intel compilers, you may use types:
INTEGER(KIND=8) for Fortran
long long int for C/C++
Note that code written this way will not work for the LP64 interface. Table 3-4 summarizes
usage of the integer types."
(bold mine)

Linking errors with mkl_intel_thread

I have implemented helper functions that would call certain Lapack and BLAS functions to calculate matrix inverses, mat-mat multiplications, and mat-vec multiplications. However, when I'm linking with the intel MKL libraries, I keep on getting the following linker errors, which I have no idea how to solve. Any suggestions would be much appreciated.

Using MKL for the first time - setup help!


I just started using Fortran and Linux a week ago, so I'm still in the progress of learning.

I'm using the Intel Fortran Compiler v11 to compile some Fortran code on quantum mechanics. I'm down to a problem where i need to solve a symmetrical n*n matrix, and someone told me to just use the "ssyev" command. This has, however, proved to be more diffifcult than I imagined!

Apparantly the "ssyev" command is from this mathpack, LAPACK, included in MKL, and to use it i got something like this:

program gaulag

(lots of stuff here)

Storage in memory of CCE format for 3d fft in C

can you please explain me how is the data stored in fourier domain for inplace 3D FFT in C? I understand that the only available format is CCE, which is not completely clearly explained in the reference manual (for 10.1 it is page 2821). I would like to know where should I find the real and complex part of coefficients and how is the data stored in memory.

Subscribe to Intel® Math Kernel Library