I have used few BLAS and LAPACK functions in my code on OpenSUSE (64-bit). The code is compiled successfully but when the linking is in process, unresolved error message is shows. I have tried to add new compiler switches but the erorr is not solved. I also ran the environtment script from tools folder at MKL directory. The output I get is:
I recently installed the Intel C++ and Fortran 11.0 compilers on my OS X machine and I noticed that this included the MKL (along with the IPP). Formerly, the MKL resided in /Library/Frameworks, but now it appears to be in /opt/intel/Compiler/11.0/054/Frameworks. My question is, can I blow away the old MKL installation in /Library/Frameworks?
I'm using PARDISO to solve a linear system originated from a Stokes problem. The matrix is real and structurally symmetric.
My first problem is that if I treat it as structurally symmetric (mtype=1) the solver fails and returns:
*** error PARDISO ( numerical_factorization) error_num= -1
*** error pardiso: zero pivot
error in PARDISO = -4
I am now using Intel MKL instead of public domain Clapack.
Can someone tell me why I get the following error when I run the program: This error doesn't come with
** On entry to DGESV , parameter number 4 had an illegal value
typedef MKL_INT integer;
integer nrhs = 1;
integer lda = m;
integer ldb = n;
ipiv = ( integer *) malloc( m*sizeof(integer) );
// A = gen_gen_matrix(m,n);
// b = gen_gen_matrix(m, nrhs);
As we know, the key to any iterative solver, e.g. GMRES, CGS, is the matrix-vector multiplication. So, I guess one of the main operation of Pardiso is to compute the matrix-vector multiplication. However, for some reason, I can not get teh explict form of the matrix to be solved, and I want to compute the matrix-vector multiplication out of the Pardiso. Therefore, I wonder whether Pardiso support this ?
Thank you !
I just started with usin IFORT and MKL, so I don't have a lot of experiences with them.
I have a code in fortran, which uses mkl for FFT. I try to compile it on my computer. I installed IFORT ver. 10.1.018 nad MKL ver. 10.0.5.025. I compile like this:
ifort -w95 -cm -O3 -align -o prog prog.f /opt/intel/mkl/10.0.5.025/lib/32/libmkl_intel.a -lpthread -Vaxlib
However, I receive an error:
/tmp/ifortSI5ZwN.o: In function `MAIN__':
prog.f:(.text+0x1465): undefined reference to `zfft2d_'
the MKL documentation states, that functions mkl_scsrmv and mkl_cspblas_scsrgemv should exist. However, they are not in mkl_spblas.h. I can only find mkl_dcsrmv and mkl_cspblas_dcsrgemv there. I would like to use single-precision floats.
Thank for your help.
I'm trying to compile a home-built code using fftw3 subroutines using MKL & ifort through wrappers on Mac OS. I built wrapper using GNU C compiler and it seems that wrapper is OK. however when I do linking of the code with external libraries I get an error:
Does anybody know how to call a library from within a function? I'm trying to convert some individual programs into functions for a larger program. When I was running the individual program, a declaration of USE GQRUL_INT and CALL GQRUL (N, QX, QW, ALPHA=ALPHA, BETAW=BETAW) to use the library was working fine. After I converted the program to a function, which is being called from a module in another file, everything to do with that library call is coming up as an error.
the solution of the spare matrix problem, we have to deal with, may fail because of some data errors. In many cases, it is helpful to know where the solver detected the singularity, e.g. this is the index of the pivot element (diagonal pivoting is assumed here).
Checking the error codes of PARADISO, the error code just says "-4" (for zero pivot, numerical factorization or iterative refinement problem).
What other information can be delivered by PARADISO? Does there exist a possibility to get the information at which pivot element the factorization failed?