I am writing some code right now and I have a placeholder with matmul which seems to work very well, but I would like to use the LAPACK dgemm implementation. I'm using gfortran right now and getting very good speeds with matmul , but I wonder if I can get better.
Current call:
C = transpose(matmul( transpose(A), B))
where A , B and C are non-square double precision matrices. I can easily write a wrapper for dgemm with the current dgemm LAPACK implementation, but I like that I can do it all as a function (instead of worrying about call for a frock coat and dealing with transpose ).
I am wondering if I compile with ifort and enable MKL , will this matmul change to MKL dgemm for me without a shell?
fortran intel lapack
drjrm3
source share