Subdividing BLACS grid

Intel® oneAPI Math Kernel Library

Ask questions and share information with other developers who use Intel® Math Kernel Library.

Subdividing BLACS grid

229 Views

I am trying to use Scalapack to calculate the eigenvalues of a large matrix using the pzheevd. If I use all processes to calculate a single matrix everything seems to work fine. But i would like to calculate different matrices in parallel:

Say I have 32 processes then I'd divide it into a 32 = 4 x 8 grid. Where I distribute my matrices across 4 processes and then calculate 8 matrices in parallel.

I tried to do this using MPI_Comm_split, but I realized that Scalapack doesn't use a communicator provided by the user. Instead it uses BLACS grids. How can I distribute my Scalapack matrices such that I have parallelism between matrices as well as within one matrix?

I couldn't find any examples and I couldn't understand how the BLACS grid is divided. Could you tell me some decent resources? Are there any alternative libraries that rely on MPI rather than PVM?

Link Copied

0 Replies

Community support is provided during standard business hours (Monday to Friday 7AM - 5PM PST). Other contact methods are available here.

Intel does not verify all solutions, including but not limited to any file transfers that may appear in this community. Accordingly, Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

For more complete information about compiler optimizations, see our Optimization Notice.