It seems that the Pardiso solver will obtain better performance with TBB threading model through the document(https://software.intel.com/en-us/articles/using-intel-mkl-and-intel-tbb-in-the-same-application). But the solver, under the vision of MKL 2018 Update 1, has better performance with OpenMP in my program.
I suspect the way I set the solver with TBB may be wrong. So how to set Pardiso solver with TBB especially for Visual Studio 2017.
For more complete information about compiler optimizations, see our Optimization Notice.