Introduction to mkl and knowledge about food products

Introduction to mkl and knowledge about food products
Introduction to Intel MKL

Intel's core mathematical function library (MKL) is a set of highly optimized and thread-safe mathematical routines and functions for high-performance engineering, scientific, and financial applications. IntelMKLThe cluster version includesScaLAPACKFast Fourier transformation with distributed memory and Linear Algebra(BLAS, LAPACKAnd Sparse Solver), fast Fourier transformation, vector mathematics(Vector Math)Supported by the random number generator.

It mainly includes:

① LAPACK (linear algebra tool linear algebra package)

② DFTs (Discrete Fourier transformation Discrete Fourier transforms)

③ VML (Vector Math Library)

④ VSL (Vector statistics Library Vector Statistical Library)

 

2. Main Functions of MKL

1) BLASAnd LAPACK

Deploy highly optimized Basic Linear Algebra routines BLAS (Basic Linear Algebra Subroutines) and Linear Algebra Package LAPACK (Linear Algebra Package) on Intel processors, which provide significant performance improvements.

 

2) ScaLAPACK

ScaLAPACK is a parallel computing package suitable for Distributed Storage MIMD parallel machines. ScaLAPACK provides a number of linear algebra solutions, which are efficient, portable, scalable, and highly reliable. It can be used to develop parallel applications based on linear algebra.

ScaLAPACKIntel? MKL implementation can provide significant performance improvements, far beyond the implementation of standard NETLIB.

 

3) PARDISO sparse matrix Calculator

The PARDISO direct sparse matrix solver is used to solve large-scale sparse linear equations. This solver is authorized by the University of Basel and is a memory-efficient software library that is easy to use and has thread security and high performance. Intel? MKL also includes the composite gradient Solver and FGMRES iterative sparse matrix solver.

 

4) Fast Fourier Transformation (FFT)

Make full use of the multidimensional FFT subprograms (from one dimension to seven dimensions) with easy-to-use new C/Fortran interfaces ). Intel? MKL supports distributed memory clusters using the same APIs and easily distributes workloads to a large number of processors, greatly improving performance. In addition, Intel? MKL also provides a series of C-language routines ("wrapper") that can simulate FFTW 2.x and 3.0 interfaces, so that the current FFTW user can convert intel? Integrate MKL into existing applications.

 

5) vector Math Library (VML)

The Vector Math Library uses computation-intensive core mathematical functions (such as power function, trigonometric function, exponential function, hyperbolic function, and logarithm function) to significantly improve the application speed.

 

6) vector statistics library-random number generator (VSL)

The Vector Statistical Library random number generator is used to accelerate the simulation, which is much higher than the scalar random number generator.

Intel MKL product introduction:

Http://software.intel.com/zh-cn/articles/intel-mkl/ (http://software.intel.com/en-us/articles/intel-mkl)

Links to Intel MKL-related documents (provide MKL-related documents ):

Http://software.intel.com/en-us/articles/intel-math-kernel-library-documentation/

Intel MKL Reference Manual (Reference Manual) (can be used to query descriptions and usage of all MKL functions ):

Http://software.intel.com/sites/products/documentation/hpc/mkl/mklman/index.htm

Compilation parameter description:-lmkl_intel_c-lmkl_intel_thread-lmkl_core-lpthread

-- To link Intel MKL, You need to include the following library files:
1: interface library functions.
2: The thread library function. (Specify the multi-threaded MKL and return the single-threaded MKL library ).
3. Optimize the core library.
4. multi-threaded support libraries.

-Lmkl_intel_c: interface library function, which is called in c mode.
-Lmkl_intel_thread: multi-threaded MKL library.
-Lmkl_core: mkl core optimization function.
-Lpthread: multi-threaded function library.