I have a computer with 4 Intel Xeon Phis installed, and I would like to get WRF running on all of them. I have replaced all calls to LAPACK functions with Intel MKL functions as described in this article:http://software.intel.com/en-us/articles/performance-hints-for-wrf-on-intel-architecture
However, I would like to enable Automatic Offload mode for MKL and for other components of WRF.
Now I read that this is done by using the environment variable MKL_MIC_ENABLE, and that it would report that it's offloading when I set OFFLOAD_REPORT = 2. However, I get no indication of WRF ever offloading anything to the MICs, and the MICs show no usage, only idleness.
What would be the best way to go about running WRF across both the host (two Intel Xeon processors) and the four MICs that I have installed?