Clik here to view.
Major regression between 11.3 Update 2 and Update 3
Hello,I'm using the QR factorization routines (?geqrf and ??mqr) extensively and it seems there is a major problem when using these routines when a workspace query is assumed (lwork = -1).Attached is...
View ArticleUbuntu mkl_set_num_threads ignored
Good afternoon!I have the following code:/* ******************************************************************************** * Copyright(C) 2004-2011 Intel Corporation. All Rights Reserved. * *...
View ArticleCan I use MKL functions with user allocated data
Hey Is there a problem using MKL with user allocated data ?I'm trying to use MKL function vcMulByConj(...)with continues memory allocated using OpenCV mat object. (with I believe it's implemented with...
View ArticleMKL Diagonal SpMV
Hi Experts,I am trying to do SpMV using the diagonal storage format. I found 2 routines that do this operation for real double-precision one-based Indexing (mkl_ddiamv, mkl_ddiagemv).I get the right...
View ArticleMKL help with mkl_sparse_s_create_coo and MKL_SPARSE_CONVERT_CSR
Hello,I am trying to use mkl_spblas and I am having some strange issues. Firstly, here is the code I am trying to compile using mkl-16.2.181 and impi-intel-5.1.3-16.0.2. MODULE testcontainssubroutine...
View Articlezgemm3m and MIC automatic off load
I have noted that zgemm will be taken by AO to the Xeon PHI, but not zgemm3m. This is a bit annoying as AO can give a significant boost to zgemm performance, but zgemm3m is faster than ZGEMM...
View ArticlePardiso example terminates when using 72 or more cpus
I studied the example cl_solver_sym_sp_0_based_c.c in cluster_sparse_solverc/source . I compiled it using:make libintel64 example=cl_solver_sym_sp_0_based_cIt runs fine . However the matrix is too...
View ArticleParameter 7 incorrect for cblas_dgemv
I have the following code: CBLAS_LAYOUT layout = CblasRowMajor; CBLAS_TRANSPOSE trans = CblasNoTrans; int m = 2; int n = 4; double a[8] = { 1,2,3,4,5,6,7,8 }; // m * n matrix double x[4] = { 1,2,3,4 };...
View ArticleClik here to view.
Wrong solution computed by distributed PARDISO with 0-based indexed BSR matrix
Hello,I'm attaching two .c files that should compute the solution of the linear system Ax = b, withA = [1, -3, 5, 1; -1, -2, 6, 10; 4, -4, 7, -5; 4, 2, 8, -1];andb = [1;1;0.75;1]The solution given by...
View ArticleMaking the cluster_sparse_solver Solve phase scalable in parallel
Hello,I have a problem with a symmetric matrix (either positive or semi-definite) for which I need to do many solutions as the time intergation of my method progresses. So I'm mostly interested in...
View ArticleDo MKL's vsl RNG contains a warm up process during initialization?
I recently read a paper http://www0.cs.ucl.ac.uk/staff/d.jones/GoodPracticeRNG.pdf in which the author pointed out how important initialization is in usage of PRNG and he suggested that a warm up is...
View ArticleChanged behavior in Pardiso, MKL 11.3.3 vs MKL 11.1.3
Hi,I recently upgraded from Compser XE 2013SP1U3 to Composer XE 2016U3, and thus from MKL 11.1.3 to MKL 11.3.3. This has caused a slightly changed behavior in the output from Pardiso in the following...
View ArticleClik here to view.
MKL PARDISO produces FPE
With the attached general matrix and rhs, PARDISO solution contains NaN.Major version: 11 Minor version: 1 Update version: 1 Product status: Product Build: n20131010 Processor optimization: Intel(R)...
View ArticleBug in zgesdd
I am using the zgesdd routine to find the singular value decomposition for a matrix, however, for certain sizes of the matrix the routine will cause an access violation (segmentation fault) and the...
View Articlequestion about the complexity to calculate just one eigenvalue
As we know that the complexity of diagonalization of a matrix should be N^3, where N is the dimension of the matrix. In MKL, there are a lot of functions that can calculate just some of the...
View ArticleClik here to view.
Segmentation fault in mkl_pds_lp64_assemble_csr_full
Hello,I'm trying to use the following file with 3 MPI processes but I end up with the following trace:(gdb) bt #0 0x00007ffff52ac38f in mkl_pds_lp64_assemble_csr_full () from...
View ArticleVSL for Convolution
I'm getting started with using MKL VSL for convolutions. For learning, I first I want to conduct a 1D convolution. I'm confused with the usage example given here:...
View ArticleWrong results from cblas_sgemm
Here is the code:float max_val(const float * vec, size_t sz) { float result = 0.0f; for (size_t i = 0; i < sz; ++i) { float val = abs(vec[i]); if (val > result) result = val; } return result; }...
View ArticleIntel MKL dll redistribution
I have a question about the license and usage possibilities of the intel mkl .dll's. Is it possible to include the (unaltered) mkl .dll's* in a (Windows) product for our customers? I understand that...
View Articlequestion about the manual of scalapack
The parameter "lwork" of the function pdsyevr can be determined by some formula.I don't understand some of the formula:why the subscript is decreased by 1,e.g., mb_ -1?what is csrc_nb_? I can not find...
View Article