With a 1-year subscription you will get access to the latest industry leading tools, premier support and updates to future releases.īy comparing the Polyhedron* benchmark of Intel® Visual Fortran 14.0 and Intel Visual Fortran 12.0 on the same hardware, you can see the dramatic performance gain by just recompiling the code.Īnd the Intel Fortran compiler provides performance leadership when compared with alternative compilers.įor math problems, achieve up to 4x higher performance on newest Intel® Core™ processors (codenamed Haswell) shown through the LINPACK Benchmark.īoth the latest compiler and Intel MKL are available in a variety of packages, including the Intel® Parallel Studio XE Family of products and Intel® Visual Fortran Composer XE. Intel tools help preserve your investment in existing development environments and code bases while helping boost application performance on today's and tomorrow's hardware.
On Windows*, they are compatible with Microsoft Visual Studio* 2008-2013 and Microsoft Windows* XP- 8.1. The latest compiler also contains features such as the following. Intel® tools support Intel® Xeon Phi™ product family on Windows-hosted machines.Additional support for Fortran90, F2003 and Fortran 2008, including co-arrays.New C++ and Fortran directives enable vectorization and offloading of execution on devices such as coprocessors or accelerators.Coarrays & synchronization constructs support parallel programming.Auto-parallelizer & directives to access SIMD instructions.More control over array data alignment (align arrayNbytes).Loop optimization directives: VECTOR, PARALLEL, SIMD.Access to Intel® AVX and Intel® AVX2 instructions (-xa or /Qxa).Additional details can be found in the reference guide. Intel® MKL has made many improvements in the last few releases. New Data Fitting functions covering one-dimensional algorithms for vector spline construction, cell or bin search, and evaluation, differentiation, and integration of the spline interpolants.New Conditional Numerical Reproducibility feature ensures identical run to run floating point results across different processors.Additional details can be found in the reference manual.
0 kommentar(er)
