Intel MKL slow

Intel MKL slow

Hi,

when I compare the performance of MKL-LAPACK with an ifc compiled version of the netlib LAPACK, MKL 6.1.1 is almost two times slower.

I use "MKL_SERIAL=yes" on a dual Xeon system running Redhat 9.
MKL is linked with
"-L/opt/intel/mkl61/lib/32 -lmkl_lapack -lmkl_ia32 -lguide -lpthread".

Is there a mistake in how I use MKL ?

2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

There have been some functions within LAPACK which have run slowly because of some issues with dlamch, the routine which determines certain double precision parameters for the machine.

What routine(s) are you calling?

Bruce

Leave a Comment

Please sign in to add a comment. Not a member? Join today