Zone des développeurs Intel® :
Open Source Innovation Program

Using Open Source Innovation Powered by Intel Tools

The Open Source Innovation Program supports developers in technical computing markets by making it a little easier for them to find and use innovation developed by others. To help find the innovators we provide this website and a series of webinars for innovators to discuss their capabilities. To use the innovation website we require the innovator provide a build script for their software and to compile and link with the relevant Intel software tools.
Learn More

Featured Innovator

Anaconda Accelerate

Anaconda Accelerate opens up the full capabilities of your GPU or multi-core processor to Python. It allows common operations like linear algebra, random-number generation, and Fourier transforms to run faster and to take advantage of multiple cores. Within Accelerate, Continuum’s Python-to-GPU compiler, NumbaPro, compiles easy-to-read Python code to multi-core and GPU architectures. Additionally, MKL Optimizations, an Intel™ Math Kernel Library-powered product, boosts the speed of NumPy, SciPy, scikit-learn, and NumExpr.

https://store.continuum.io/cshop/accelerate/

MKL Optimizations
Developed specifically for science, engineering, and financial computations, Continuum has packaged Intel® MKL-powered binary versions of some of the most popular numerical/scientific Python libraries into MKL Optimizations for improved performance. MKL Optimizations includes speed-boosted NumPy, SciPy, scikit-learn, and NumExpr; MKL packaging with redistributable binaries for easy access to the MKL runtime library; and Python bindings to the low level MKL service functions, which allow for the modification of the number of threads being used during runtime.

https://store.continuum.io/cshop/mkl-optimizations/

Benchmarks

More information about the machines on which the benchmarks were run can be found on the MKL Optimizations page of the Continuum site (https://store.continuum.io/cshop/mkl-optimizations).

About Continuum Analytics/Capabilities

Continuum’s Python-based, data analytics products and consulting services empower organizations to analyze, manage and visualize big data – turning massive datasets into actionable insights and business value. Built on proven, open-source offerings and integrated within existing IT environments, Continuum Analytics’ tools allow organizations to make critical business decisions based on their data quickly, easily, inexpensively, and flexibly.

Contact Information

Continuum Analytics, Inc.
info@continuum.io
www.continuum.io

Using Open Source Innovation Powered by Intel Tools

The Open Source Innovation Program supports developers in technical computing markets by assisting them with users or customers demanding that, for certain problems, execution speed must be increased three to ten times beyond single workstation performance.

To meet the demand, developers are looking to deliver functionality that will enable future execution on a wide range of computers – i.e. laptops, workstations and clusters, and plan to increase the application of parallelism to meet performance demands. These developers are using modern high-level languages (i.e. – C++, Python, Java, SciPy/NumPy) to develop the solutions.

On technical fronts, rapid advances in computational platforms have complicated the task of effectively leveraging new technology. Developers want to avoid delivering software that is often out-of-date as soon as it is released.

If you are dealing with the challenges of exponential data growth that mean newer and more complex numerical methods for scientific analysis, read on.

An approach to stay ahead of the “technology Curve”

A community of experts is very good at delivering software that can execute on shared memory laptops, desktops and servers, distributed memory clusters and today’s hybrid systems that include attached processors. These experts around the world make their innovation available to the developer community via source that is distributed. And Intel works with these developers to ensure that performance of their software is fully exposed on Intel-based platforms.

How Intel Helps

Intel offers a full suite of development tools (Cluster Studio XE) that help with maximizing performance in a minimum amount of time for software written in C, C++ and Fortran. We provide a number of libraries (MKL, TBB, MPI) to enable sustainable performance as the Intel architecture evolves. Finally, we work with a number of open source and application suppliers to enable their software to be performance tuned for evolving Intel architectures.

We are consolidating the information on the work that Intel has done and providing a vehicle for open source innovators to communicate their capabilities.

The vehicle consists of two parts – a website listing open source innovation and a webinar series for open source innovators to market their innovation to potential users

Initially, this website provides a list of open source applications and linkable solvers that

  • Have been fully or partially tested by Intel application engineers
  • Provide a standard build capability for open source and Intel software tools
  • Plan to scale forward as Intel architecture evolves.

As the number of innovators grows, the testing by Intel application engineers will reduce and the self certification for a standard build capability with increase.

Along with the innovators, Intel will provide a list of webinars that will more formally describe the innovation and the performance delivered by Intel cluster tools.

Anaconda Accelerate

Anaconda Accelerate opens up the full capabilities of your GPU or multi-core processor to Python. It allows common operations like linear algebra, random-number generation, and Fourier transforms to run faster and to take advantage of multiple cores. Within Accelerate, Continuum’s Python-to-GPU compiler, NumbaPro, compiles easy-to-read Python code to multi-core and GPU architectures. Additionally, MKL Optimizations, an Intel™ Math Kernel Library-powered product, boosts the speed of NumPy, SciPy, scikit-learn, and NumExpr.

https://store.continuum.io/cshop/accelerate/

Benchmarks

More information about the machines on which the benchmarks were run can be found on the MKL Optimizations page of the Continuum site (https://store.continuum.io/cshop/mkl-optimizations).


Trilinos

The Trilinos Open Source software project has been developed by Sandia National Labs (along with additional open source developers) to create algorithms and tools for use within an object-oriented software framework in order to create solutions for large-scale, complex multi-physics engineering and scientific problems.

Trilinos Performance Comparison


VisIt

VisIt is a free interactive parallel visualization and graphical analysis tool for viewing scientific data on Unix and PC platforms. Users can quickly generate visualizations from their data, animate them through time, manipulate them, and save the resulting images for presentations. VisIt contains a rich set of visualization features so that you can view your data in a variety of ways. It can be used to visualize scalar and vector fields defined on two- and three-dimensional (2D and 3D) structured and unstructured meshes. VisIt was designed to handle very large data set sizes in the terascale range and yet can also handle small data sets in the kilobyte range.

VisIt Gallery

Past Webinars