More Resources Intel C++ and Fortran Compilers Intel Math Kernel Library (Intel MKL)

The libraries that previously were available separately still are, but the professional edition combines them into a single package using one serial number at a bundled price. Now, as a developer, you have a broader set of tools available when developing threaded applications, including optimized threading templates, math processing and multimedia support.

This release offers numerous compiler and Math Kernel Library (MKL) improvements and combines several key elements including the compiler, math processing, C++ templates for parallelism and multimedia libraries (the latter two only for C++). Intel C++ and Fortran Compiler Professional Editions offer advanced optimization, multi-threading and processor support, including automatic processor dispatch, vectorization, auto-parallelization, OpenMP, data prefetching, and loop unrolling in addition to the previously mentioned templates and libraries.

For multithreaded app development, two products currently exist from which you can choose. The version for Mac OS X, which was introduced last year, combined the compiler with libraries, and now the 10.0 release expands this model to all supported operating systems. Previously, the Intel Fortran Compiler Professional Edition v9.1 included the IMSL library. Now with the 10.0 version, Intel Fortran Compiler Professional Edition + IMSL, you can still purchase the compiler with the IMSL library but it also includes the Intel Math Kernel Library.

New Compiler Features

The Intel C++ Compiler Professional Edition for Windows, Linux, and Mac OS X bundle the compiler with Intel Math Kernel Library, Intel Integrated Performance Primitives and Intel Threading Building Blocks. Intel Fortran Compiler Professional Edition for Windows, Linux and Mac OS X bundles the compiler with Intel Math Kernel Library. Compaq Visual Fortran users will be glad to know that the Intel Visual Fortran Compiler for Windows now also includes the Microsoft Visual Studio development environment.

Better Performance and Threading
The new Parallel/Loop Optimizer provides improved optimization reports for you heavy-duty performance users. The latest compilers provide better performance for computationally-intensive applications, such as graphics/digital media, financial modeling and high-performance computing for threaded and non-threaded applications. Intel's new High Performance Parallel Optimizer (HPO), can improve your ability to analyze, optimize and parallelize more loop nests.

Improved Optimization in C++
Exception handling and in-class hierarchy analysis have also been improved. Intel has improved the compilers' ability to optimize in the presence of C++ exception handling, analyzing, and optimizing C++ class hierarchies.

Security Checking and Diagnostics
The latest compiler version provides GNU Mudflap, Static Verifier for buffer overflow, security, and OpenMP API verification. You now have the ability to create code that is less susceptible to security vulnerabilities, such as a buffer overflow. These diagnostics can help both novice and expert users detect common coding errors, including initialized variables, mismatched dummy and actual arguments, and OpenMP API coding issues.

Stand-alone Visual Fortran on Windows
With the Intel Fortran Compiler, Windows users no longer need to purchase Microsoft Visual Studio 2005 separately. This completes the migration for Compaq Visual Fortran users by providing a significant feature and performance improvement in a standalone solution.

64-bit Mac OS X Support
You now have the ability to create apps that take advantage of addressing and performance capabilities enabled by the Intel 64 processors used in the latest Apple Mac OS X systems.

Windows Vista and Visual Studio.Net 2005 Support
You now have the ability to develop applications that work on the latest operating system and Integrated Development Environment (IDE) from Microsoft. Intel's Visual Studio 2005 support now includes a seamless use of property pages, improving the compatibility with your existing VS 2005 projects files.

New Math Kernel Library (Intel MKL) Features

The Intel Math Kernel Library is a set of mathematical functions for engineering, scientific and financial applications requiring maximum performance. Intel MKL functions are highly optimized and multi-core ready, meaning they are extensively threaded and thread-safe. The Cluster Edition of Intel MKL provides support for distributed memory parallel processor systems, enabling extremely high levels of parallelization and performance.

MKL provides a variety of math routines for high-performance computing, such as linear algebra like BLAS, LAPACK, ScaLAPACK, and Sparse Solver. Other math routines include Fast Fourier Transforms, Vector Math and Random Number Generators.

Optimization Reports
You now can use Intel's optimization report diagnostics, using advanced optimizations, to help the compiler more effectively tune your applications.

Optimizations for Latest Intel Processors
The latest MKL provides optimizations for the latest Intel processors including Quad-core Clovertown (Xeon 5300 series) and Dual-core Woodcrest (Xeon 5100 series) CPUs. This allows you to take advantage of the latest Core2 architecture features for improved performance on these platforms. Further, new threading in vector math functions allows you to benefit from the power of the newest Intel Dual- and Quad-core architectures.

New Data Compression Functions and Performance
New data compression functions and code samples provide full compatibility with zlib and libbzip2. You now have the ability to incorporate industry-standard data compression using ready to use templates. Further, data compression is faster by simply linking the updated library.

New bzip2 Functions and gzip-compatible Command Line Utility Code Sample
The latest MKL provides data compression routines targeting enterprise developers. Integrated Performance Primitives (Intel IPP) implementations now are faster than generic versions you may be using currently. IPP is a software library of highly-optimized functions for multimedia, audio, video, speech and computer vision, as well as image and signal processing.

Support for New VC-1 and H.264 High Profile Video codecs
MKL has the ability to incorporate ready-made templates that can render video using the latest video codecs. Intel IPP has new Unified Media C++ Classes (UMC) for media application development, including H.264 codecs, as well as optimized functions covering 14 domains and downloadable code samples.

LAPACK 3.1 Support
MKL is compliant with the latest LAPACK specification, defined in the first quarter of 2006. The LAPACK math domain makes up 69 percent of MKL users.

Automatic Grain Size for Better Parallel Algorithms
Intel MKL provides automatic grain size for better parallel algorithms. It automatically subdivides a task to get the best performance on multi-core processors.

Support for Microsoft Vista and Linux Distributions
You now have the ability to deploy applications on the latest operating systems, including Microsoft Vista and Linux distributions. It also provides full, native 64-bit support for Mac OS X. You also now can utilize TBB for 64 bits on Tiger and the upcoming Leopard Mac operating systems.

Intel TBB Academic Instructor Program
Academic instructors have not been forgotten. You can receive a free Intel TBB academic license. Just complete the survey before August 31, 2007.

Regardless of your platform of choice, you now have a bundle of developer tools that can help increase your productivity when developing multithreaded applications. New enhancements in Intel's Fortran and C++ compilers and Intel MKL can help you develop applications with maximum performance and efficiency at a single, bundled price.