CppTensority

Prerequisites

• OpenBLAS

How to speed it up

If you don't mind the portability issue, you may try to opt it by working on the cblas_dgemm() using:

• MKL on a Intel machine, or even
• GPU computing lib, for example, cuBLAS
  • cublasGemmEx may draw your interests and you would have to convert the types and prepare (set up) the matrix beforehand