Master

A sample code for performing Sparse Matrix-Matrix Multiplication (SpGEMM) using MKL has been added.

How are the following issues solved?

1. The line "sparse_linear_algebra/spgemm/mkl_shmem" should be added into dwarfs_to_build.json.

2. The following compiler.json file must be used:

   { "Intel": { "cc": "icc"} }