Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.
6977 Discussions

MKL Scalapack uses much more RAM than NETLIB version

Razinkovas__Lukas
Beginner
‎11-17-2019 02:14 AM
483 Views

Hello,

I really enjoy using MKL libraries since they are much faster on our 36 core Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz mashine. However our scientific research requires diagonalization of very large matrices which consume a lot of memory and MKL scalapack uses much more memory compared to NETLIB version. I am attaching graph where i show memory consumption of our code using intel and netlib libraries. Diagonalisation (pdsyevr) is performed between vertical lines. One can notice sudden leap of used memory in the middle of diagonalisation procedure in MKL version. Is it possible somehow to get rid of this behavior since it drains all RAM.

used_memory.png

Sincerely,

Lukas

 

0 Kudos

Link Copied

1 Reply
Gennady_F_Intel
Moderator
‎11-19-2019 10:06 PM
483 Views

What is the problem size and how many mpi processes you are running this case?

0 Kudos
Copy link
Reply

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.