The MKL_DYNAMIC environment variable enables Intel MKL to dynamically change the number of threads.
The default value of MKL_DYNAMIC is TRUE, regardless of OMP_DYNAMIC, whose default value may be FALSE.
When MKL_DYNAMIC is TRUE, Intel MKL tries to use what it considers the best number of threads, up to the maximum number you specify.
For example, MKL_DYNAMIC set to TRUE enables optimal choice of the number of threads in the following cases:
If the requested number of threads exceeds the number of physical cores (perhaps because of hyper-threading), and MKL_DYNAMIC is not changed from its default value of TRUE, Intel MKL will scale down the number of threads to the number of physical cores.
If you are able to detect the presence of MPI, but cannot determine if it has been called in a thread-safe mode (it is impossible to detect this with MPICH 1.2.x, for instance), and MKL_DYNAMIC has not been changed from its default value of TRUE, Intel MKL will run one thread.
When MKL_DYNAMIC is FALSE, Intel MKL tries not to deviate from the number of threads the user requested. However, setting MKL_DYNAMIC=FALSE does not ensure that Intel MKL will use the number of threads that you request. The library may have no choice on this number for such reasons as system resources. Additionally, the library may examine the problem and use a different number of threads than the value suggested. For example, if you attempt to do a size one matrix-matrix multiply across eight threads, the library may instead choose to use only one thread because it is impractical to use eight threads in this event.
Note also that if Intel MKL is called in a parallel region, it will use only one thread by default. If you want the library to use nested parallelism, and the thread within a parallel region is compiled with the same OpenMP compiler as Intel MKL is using, you may experiment with setting MKL_DYNAMIC to FALSE and manually increasing the number of threads.
In general, set MKL_DYNAMIC to FALSE only under circumstances that Intel MKL is unable to detect, for example, to use nested parallelism where the library is already called from a parallel section.