What’s New in Intel® Composer XE 2013 SP1

Intel® Composer XE 2013 SP1 includes Intel® Compiler 14.0 among other components. The list below summarizes the features and enhancement highlights in Intel® Compiler 14.0 that are pertinent to those programming for Intel® Xeon Phi™ coprocessors:

·       Support for the new Intel® Xeon Phi™ Coprocessor, Code Named Knights Landing  

·       New Data Alignment features (C++ align_value attribute and Fortran ASSUME directive)

·       Automatic detection of new Intel® MPSS 3.1 release

·       Indirect prefetching enabled using pragmas for 14.0

·       Enhanced OpenMP* 4.0 simd-vectorization and RC2 Offload

·       Enhanced Coarray Fortran on Intel® Xeon Phi™ Coprocessor

·       Enhanced Integrated vectorization report and program listing

·       Enhanced offload via shared keyword in Windows*

·       Enhanced shared use of Intel® Xeon Phi™ Coprocessors

·       Expanded support for high-performance math in scalar code

·       Enhanced memory prefetching for direct memory references

·       Improved offload performance for persistent data

·       Improved vectorization and code-generation

 

Below is more detail on the change found in Intel® Composer XE 2013 SP1 since the latest composer release (Intel® Composer XE 2013). Please refer to Intel® C++ Composer XE 2013 SP1 Release Notes or Intel® Fortran Composer XE 2013 SP1 Release Notes for more information.

 

Intel® C++ Composer XE 2013 SP1 for Linux*

·       Online installation

·       Graphical User Interface installation

·       Intel® C++ Compiler XE 14.0.0

·       GDB* is based on GDB 7.5, and offered other new features: support for Intel® MIC Architecture, Intel® Parallel Debug Extension, Trace enhancements, Intel® Pointer Checker, and Intel® Transaction Synchronization Extension

·       Fedora* 18 and 19 are now supported

·       Ubuntu* 13.04 and Debian* 7.0 are now supported

·       Support for the following versions of Linux distributions has been dropped:

o   Fedora* 17

o   Ubuntu* 11.10

o   Pardus* 2011.2

·       Features from C++11 (-std=c++11)

·       Partial OpenMP* 4.0 RC2 (Public Review Release Candidate 2) support

·       Intel® Cilk™ Plus changes: elemental function implementation has changed to be more compatible with other vector function implementations in gcc and OpenMP*, added new multiply reducer, added three new array notation reduction intrinsics.

·       DWARF V4 support

·       __INTEL_COMPILER_UPDATE predefined macro

·       Pointer type alignment qualifiers

·       Variable definition attributes to avoid false sharing

·       -mtune performance tuning option

·       Using offload code in shared libraries requires main program to be linked with –

offload=mandatory or –offload=optional option

·       -openmp-offload/-openmp-simd options added for controlling the enabling/disabling of specific OpenMP* 4.0 features independently of other OpenMP features

·       __GXX_EXPERIMENTAL_CXX0X__ Macro Not Supported

·       -xATOM_SSE4.2 option added to support Silvermont microarchitecture

·       Intel® Math Kernel Library 11.1

·       Intel® Integrated Performance Primitives 8.0 update 1

·       Intel® Threading Building Blocks 4.2

 

Intel® C++ Composer XE 2013 SP1 for Windows*

·       Online installation

·       Intel® C++ Compiler XE 14.0.0

·       Support for Intel® Many Integrated Core Architecture (Intel® MIC Architecture)

·       Preview Support for Intel® Graphics Technology

·       Features from C++11 (-std=c++11)

·       Partial OpenMP* 4.0 RC2 support

·       Intel® Cilk™ Plus changes

·       __INTEL_COMPILER_UPDATE predefined macro

·       Pointer type alignment qualifiers

·       Variable definition attributes to avoid false sharing

·       Using offload code in shared libraries requires main program to be linked with

offload=mandatory or –offload=optional option

·       Limitations of _Cilk_shared

·       /Qopenmp-offload and /Qopenmp-simd options added for controlling the

·       enabling/disabling of specific OpenMP* 4.0 features independently of other OpenMP* features

·       /QxATOM_SSE4.2 option added to support Silvermont microarchitecture

·       Intel® Debugging Extension 1.0 for Intel® MIC Architecture

·       Intel® Math Kernel Library updated to version 11.1

·       Intel® Integrated Performance Primitives updated to version 8.0 update 1

·       Intel® Threading Building Blocks updated to version 4.2

 

Intel® Fortran Composer XE 2013 SP1 for Linux*

·       Intel® Fortran Compiler XE 14.0.0

o   Support added for running coarray applications on Intel® Xeon Phi™ coprocessors

·       Intel® Debugger support deprecated

·       Intel® Math Kernel Library updated to version 11.1

·       Enhanced GDB debugger provided

·       A Graphical User Interface form of the installer is provided as an option

·       An Online form of the installer, where only required components are downloaded, is provided as an option

·       Support for Ubuntu* 13.04, Fedora* 18, 19, and Debian* 7.0 added.

·       Support for the following Linux distributions has been dropped:

o   Fedora* 17

o   Ubuntu* 11.10

o   Pardus* 2011.2

 

Intel® Visual Fortran Composer XE 2013 SP1 for Windows*

·       Intel® Fortran Compiler XE 14.0.0

o   Support added for developing applications that use Intel® Xeon Phi™ coprocessors

o   Intel® Debugger Extension for Intel® Manycore Integrated Core Architecture (Intel® MIC Architecture) provides debugging applications that use Intel® MIC Architecture from Microsoft Visual Studio 2012*

·       Intel® Math Kernel Library updated to version 11.1

·       An online version of the installer, where only required components are downloaded, is provided as an option

 

 

Reportez-vous à notre Notice d'optimisation pour plus d'informations sur les choix et l'optimisation des performances dans les produits logiciels Intel.