Are symmetric and packed storage routines threaded?

Discussion created by golovashkin on Apr 14, 2010

I am curious whether matrix-vector and rank-1 and rank-2 updates for symmetric and packed storage are multithreaded.

dspr2_  dsyr2_

dsymv_  dspmv_

Experiments suggest that they are not:  Changes in OMP_NUM_THREADS (exported var) had no effects on the resultant execution time.   Am I correct?   If yes, any plans to make those threaded?