I am trying to solve a system with `cluster_sparse_solver`, and am getting unexpected results. https://gist.github.com/ivan-krukov/157372d9a55db244c4b4
In my case, given a sparse matrix `A`, we want to solve `Ax=e`, where `e` is the first column of the identity matrix with the appropriate size (A column where the first entry is 1, rest are zeros).
I am doing development on a 24-core machine (E5-2697-v2). When I launch a single DGEMM where the matrices are large (m=n=k=15,000), the performance improves as I increase the number of threads used, which is expected. For reference, I get about 467 GFLOPs/sec using 24 cores.
I am feeding pardiso a diagonal matrix with several zeros on the main diagonal. It is returning error=-1, "input inconsistent", when I was expecting -7 for "diagonal matrix is singular".
What exactly does "input inconsistent" mean?
I build sucessfully R 3.2.2 with Intel MKL and ICC. Now I am wondering if library linking are indeed correct. Here is the output:
I have a problem with setting the input parameters at MKL FEAST algorithm library.
(I currently use MKL 11.3 version and Intel parallel studio xe 2016 for c++ at Linux)
I was trying to use the dfeast_syev function, so I set the extended eigensolver input parameters by using the array fpm, as written in the reference manual.
When I modify the fpm values to change the options for dfeast_syev function, I get errors for setting illegal number in fpm.
In many mkl_sparse_XXX routines (version 11.3), there is a a parameter of type matrix_descr, just like :
sparse_status_t mkl_sparse_s_trsv (sparse_operation_t operation, float alpha, const sparse_matrix_t A, struct matrix_descr descr, const float *x, float *y);
The explanation of this parameter in manual is:
descr : Structure specifying sparse matrix properties. ......But it is confusing. Following is the code snippet in MKL example sparse_trsv.c,