Intel Math Kernel Library (MKL)

de Intel

Overview

Intel® Math Kernel Library (Intel® MKL) is composed of highly optimized functions for math, engineering, scientific, and financial applications that require high performance on Intel platforms. The functional areas of the library include linear algebra consisting of LAPACK and BLAS, discrete Fourier Transforms (DFT), vector transcendental functions (vector math library/VML) and vector statistical library (VSL) functions.

Features and Benefits

• High Performance Floating-Point Math
• Linear algebra with LAPACK (solvers and eigensolvers) plus BLAS Levels 1, 2, 3 (vector, vector-matrix, matrix-matrix) operations
• 1D, and up to 7D Discrete Fourier Transforms with complex-to-complex, real-to-complex, complex-to-real and conjugate even types
• Vector math library (VML) provides a set of highly optimized implementations of certain computationally expensive core mathematical functions, including power, trigonometric, exponential and hyperbolic, providing vector transcendental functions offering speed over scalar versions
• Vector statistical library (VSL) offers vectorized random number generators for improved performance
• Multi-threading support using OpenMP plus a fully thread-safe library
• High performance on Intel® Pentium® 4, Intel® Xeon?, Itanium® 2 processors, and Pentium M ? a processor component of Intel® Centrino? mobile technology¹
• Trial version is available
  

What?s New in the Intel Math Kernel Library

• Intel Math Kernel Library continues to excel by delivering improved performance and additional numerics functionality to meet the demands of today?s applications and simulation workloads on Intel processor-based desktops, workstations and servers.
• The DFT extends the FFT coverage of Intel Math Kernel Library, providing one-dimensional and multi-dimensional routines (up to the order of 7) with transform lengths of other than powers of 2 (mixed radix support).
• The vector statistical library offers high performance random number generators that are hand assembled, tuned, and vectorized. These pseudorandom number-generator subroutines provide both basic continuous and discrete distributions.
  

Intel Math Kernel Library (Intel MKL) Features Optimize your Code

• Focus your development time on building value-add features using the highly-optimized library functions of Intel Math Kernel Library, to help you get your optimized applications to market faster.
• Write once to the Intel MKL application programming interface (API) and yield performance benefits across the range of Intel microprocessors, saving you time.
• Parallelism Built-in ? Many functions in Intel Math Kernel Library are threaded, providing the advantages of parallelism, including excellent scaling for applications with none of the work.
  

Performance Improvements: Linear Algebra and DGEMM

Matrix multiplication is a cubic operation; doubling the sizes of the matrices to be multiplied increases the number of operations by a factor of 8. Double-precision general matrix-matrix multiply (DGEMM) is the workhorse routine for dense linear algebra. In many applications that are heavily dependent on the solution of large systems of equations, the performance of properly-written solvers depends directly on DGEMM performance.

Powerful Computing on Your Desktop

The various functional areas of Intel MKL (LAPACK, BLAS, DFTs, VML and VSL) provide optimized performance for the latest features and capabilities of the Pentium 4, Intel Xeon, Intel Itanium 2 processors and Pentium M ? a processor component of Intel Centrino mobile technology1. Many functions in Intel MKL have been threaded to provide greater performance on symmetric multiprocessing (SMP) systems. This provides the advantages of parallelism, including excellent scaling for applications, with none of the work.

Applications Optimized for Intel® Processor-based Systems

Intel Math Kernel Library is compatible with Windows and Linux environments, including Microsoft Windows 2000, Millennium Edition, and XP. Linux support includes Linux distributions with 2.4x kernels or later. Intel MKL runs with Intel Fortran Compiler 6.0 or later, Compaq Visual Fortran 6.0, Intel® C++ Compiler 6.0 or later, and Microsoft Visual C++ Compiler 6.0 or later. Whether you program in Fortran or C, on Windows or Linux, Intel MKL optimizes the performance. Both Fortran and C interfaces exist for the basic linear algebra subroutines (BLAS); the C interface for the BLAS is also referred to as CBLAS. The LAPACK functions are available with a Fortran interface. Fortran and C interfaces are available for all of the Discrete Fourier Transform functions, the vector math library (VML) functions and the vector statistical library (VSL) functions.

Intel® Premier Support Included with Every License

Every purchase of an Intel® Software Development Product includes a year of support services, which provides access to Intel® Premier Support and all product updates during that time. Intel Premier Support gives you online access to technical notes, application notes, and documentation. Install the product, and then register to get support and product update information.

Hardware Requirements

• Pentium 4, Intel Xeon, Itanium 2 processors, or Pentium M ? a processor component of Intel Centrino mobile technology
• 32 MB of RAM (large problems may require more memory)
  

Software Requirements

• Microsoft Windows 2000, XP, ME, or Windows NT 4.0
• Linux versions: Kernel v 2.4 or later
• Intel Fortran Compiler 6.0 or late
• Compaq Visual Fortran 6.0
• Intel C++ Compiler 6.0 or later
• Microsoft Visual C++ Compiler 6.0 or later