How to set affinity of threads spawned by MKL?

I'm trying to do some optimization on a NUMA system at the moment and have run into trouble with regards to thread affinity.

The essence of the code I'm using can be seen below:

      success = SETENVQQ("OMP_PLACES={0,1,2,3,4,5,6,7},{8,9,10,11,12,13,14,15},{16,17,18,19,20,21,22,23},{24,25,26,27,28,29,30,31},{32,33,34,35,36,37,38,39},{40,41,42,43,44,45,46,47},{48,49,50,51,52,53,54,55},{56,57,58,59,60,61,62,63}")
      NUMA=4
      ThreadsPrNuma=2
      call omp_set_dynamic(0)
      call mkl_set_dynamic(0)
      call omp_set_nested(1)
      call omp_set_num_threads(NUMA)
      call mkl_set_num_threads(NUMA)
      !$OMP PARALLEL DEFAULT(SHARED) PRIVATE(k) proc_bind(spread)
      !$OMP DO SCHEDULE(STATIC)
      do k=1,NUMA
        !$OMP PARALLEL DEFAULT(SHARED) PRIVATE(i) proc_bind(master)
        !$OMP END PARALLEL
        call omp_set_num_threads(ThreadsPrNuma)
        call mkl_set_num_threads(ThreadsPrNuma)
        call MKL_routine_to_run_parallel_on_each_NUMA_node()
      end do
      !$OMP END DO
      !$OMP END PARALLEL

So the idea is that I have I want to do an outer paralization over each NUMA node - proc_bind(spread).

And then I want each of these threads to have an inner paralization over the number of processors in that NUMA node, the problem is that the MKL routine doesn't seem to be respecting the proc_bind(master) I set.

Anyone got any idea how to do such a thing?

Cheers

Tue