Intel MKL is threaded extensively for different domains. The threaded function list includes:
Direct sparse solver
DSS Interface Routines
Orthogonal factorization, computational routines:
Singular Value Decomposition, computational routines: *gebrd, *bdsqr
Symmetric Eigenvalue Problems, computational routines:
Level2 BLAS: *gemv, *trmv, dsyr/ssyr, dsyr2/ssyr2, dsymv/ssymv
All Level 3 BLAS and all Sparse BLAS routines are threaded except Level 2 triangular solvers
DFT transform are threaded except the following:
1)Real 1D are not threaded
Note 1: A number of other LAPACK routines, which are based on threaded LAPACK or BLAS routines, make effective use of parallelism: *gesv, *posv, *gels, *gesvd,*syev, *heev, etc.
Note 2: Level1 BLAS and Level2 BLAS are threaded only for:
1) Intel® 64 architecture
2) Intel® CoreTM2 Duo and Intel® CoreTM i7 processors
Note 3: all data provided into this article are relevant to the latest MKL.
Note 4: The functions are not threaded when small vector/matrix size. For example, VML function are expected to be threaded for vector length > 16K.