Вернуться к обычному виду

Software development toolkits

Intel Cluster Studio

Intel Cluster Studio XE 2013 
Intel Cluster Studio XE toolkit provides a comprehensive set of parallel programming standards driven by C/C++ and Fortran development tools and programming models which enable software developers to efficiently develop, analyze, and optimize HPC applications to scale forward, scale faster, and boost performance for IA-compatible processors.

Intel Cluster Studio XE 2013 includes the software development tools: 
  • Intel C, C++ and Fortran compilers.
  • Intel MPI Library – highly scalable and interconnect independent low latency MPI library. 
  • Intel MPI Benchmarks - an open source set of MPI and cluster benchmark kernels.
  • Intel Trace Analyzer and Collector – MPI communications performance profiler.
  • Intel Math Kernel Library (MKL) and Intel Integrated Performance Primitives (IPP) – performance libraries for math and multimedia. 
  • Intel Threading Building Blocks (TBB) and Intel Cilk Plus – parallel programming models based on threading. 
  • Intel Advisor XE – threading assistant for C/C++/C#/Fortran applications using threadbased parallelism on the master node of a cluster. 
  • Intel VTune Amplifier XE – performance & thread profiler is MPI enabled on every node. 
  • Intel Inspector XE – memory and thread checker is MPI enabled on every node. 

View the Release Notes.

Intel Parallel Studio

Intel Parallel Studio XE 2015 Professional Edition 

Intel Parallel Studio XE toolkit simplifies the design, development, debug high-performance applications that use the capabilities of modern IA-compatible processors and coprocessors

Intel Parallel Studio XE 2015 Professional Edition includes the software development tools:
  • Intel C/C++/Fortran compilers.
  • Intel Math Kernel Library (MKL) – highly vectorized and threaded routines for Linear Algebra, Fast Fourier Transforms (FFT), Vector Math and Statistics functions.
  • Intel Integrated Performance Primitives (IPP) – extensive software library for media and data processing using SIMD and AVX instruction sets.
  • Intel Threading Building Blocks (TBB) – widely used С/C++ template library for task parallelism.
  • Intel Advisor XE is a threading design and prototyping tool for C/C++/C#/Fortran software architects. 
  • Intel VTune Amplifier XE – performance profiler (C++/C#/Fortran/Java).
  • Intel Inspector XE is an easy to use memory and threading error debugger for C/C++/C#/Fortran applications.
  • Intel Cilk Plus – extension to the C/C++ languages to support data and task parallelism.
View the Release Notes.

Intel Manycore Platform Software Stack

Intel Manycore Platform Software Stack (MPSS)  

Intel MPSS – the user- and system-level software that allows programs to run on and communicate with the Intel Xeon Phi coprocessor. 

The Linux software for the Intel Xeon Phi coprocessor consists of a number of components
  • Device Driver: At the bottom of the software stack in kernel space is the Intel Xeon Phi coprocessor device driver. The device driver is responsible for managing device initialization and communication between the host and target devices.

  • Libraries: The libraries live on top of the device driver in user and system space. The libraries provide basic card management capabilities such as enumeration of cards in a system, buffer management, and host-to-card communication. The libraries also provide higher-level functionality such as loading and unloading executables onto the coprocessor, invoking functions from the executables on the card, and providing a two-way notification mechanism between host and card. The libraries are responsible for buffer management and communication over the PCIe bus.

  • Tools: Various tools that help maintain the software stack. 

  • Card OS (uOS): The Linux-based operating system running on the Intel Xeon Phi coprocessor.

View the Intel MPSS User's Guide and Intel Xeon Phi Coprocessor Developer's Quick Guide.


NVidia CUDA is a parallel computing platform and application programming interface (API) model created by NVidia. It allows software developers and software engineers to use a CUDA-enabled graphics processing unit (GPU) for general purpose processing — an approach termed GPGPU (General-Purpose computing on Graphics Processing Units). 

The CUDA platform is a software layer that gives direct access to the GPU's virtual instruction set and parallel computational elements, for the execution of compute kernels.

The CUDA platform is designed to work with programming languages such as C/C++ and FortranThis accessibility makes it easier for specialists in parallel programming to use GPU resources, in contrast to prior APIs like Direct3D and OpenGL, which required advanced skills in graphics programming. Also, CUDA supports programming frameworks such as OpenACC and OpenCL

NVidia CUDA has been widely deployed through hundreds of applications.

Start accelerating your application by visiting the Getting Started Page.

View NVidia CUDA Toolkit CUDA Developer Zone, CUDA Toolkit DocumentationCUDA Toolkit Archive and Release Notes of current version.