|
contents | software | |||||||
| AMD Announces AMD Core Math Library 4.0 AMD has announced significant updates to its AMD Core Math Library (ACML) designed to provide software developers with highly optimized, multi-threaded math routines used by scientific, engineering, business and consumer applications that require maximum performance. Designed for developers and mathematical programmers seeking additional support for multi-threading and complex high performance computing projects, ACML offers support for all major x86 operating environments including Windows, Linux, and Solaris. ACML can be downloaded at no charge from AMD Developer Central, an on-line destination for developers seeking support for multi-threading software development projects or x86-based computing projects that demand highly-optimized code. ACML is a set of numerical routines tuned specifically for AMD64 processors designed to increase support for linear programming, an optimizing technique for mathematically intensive code, in multi-core processing environments. Enhancements to the new version of ACML include an update to the Linear Algebra Package (LAPACK) and increased optimizations for Quad-Core AMD Processors, including Quad-Core AMD Opteron Processors and the upcoming AMD Phenom quad-core processor. ACML is specifically designed to use key components of next-generation AMD processors, including Streaming SIMD Extensions (SSE), and to create highly threaded code that can help developers to better utilize multi-core processor technology. ACML consists of the following features: - A full implementation of Level 1, 2 and 3 Basic Linear Algebra Subroutines (BLAS), with key routines optimized for high performance on AMD Opteron processors. - A full suite of Linear Algebra (LAPACK) routines. As well as taking advantage of the highly-tuned BLAS kernels, a key set of LAPACK routines has been further optimized to achieve considerably higher performance than standard LAPACK implementations. - A comprehensive suite of Fast Fourier Transforms (FFTs) in single-, double-, single-complex and double-complex data types. - Fast scalar, vector, and array math transcendental library routines optimized for high performance on AMD Opteron processors. - Random Number Generators in both single- and double-precision. ACML is designed to be used with a variety of popular x86 C/C++ and Fortran compilers, including GCC and GFortran (The GNU Compiler Collection), PGI, Pathscale, Sun, and Intel, providing developers with freedom of choice when creating highly-optimized, multi-threaded code. With the tuned implementations of industry-standard math libraries and other frequently used scientific subroutines, ACML enables developers to accelerate complex programming efforts such as weather modeling, finite element analysis, computational fluid dynamics, financial analysis, and oil and gas applications. First made available in 2003 to deliver an enhanced development tool and provide developers with a core set of AMD64 mathematical functions, ACML was launched in collaboration with the Numerical Algorithms Group (NAG). Over the past four years, AMD has evolved ACML to assist with today's most complex mathematical problems. ACML is being used by a growing community of ISVs, corporate developers, and solution providers who are creating optimized software specifically for AMD multi-core processor technology. write your comments about the article :: © 2007 Computing News :: home page |