Home > News > Members and Partners News > INTEL

INTEL News

December 16, 2015 - 21 new projects have been awarded by the EU in 2015 in the  H2020 FETHPC program. The READEX project (exploitation of application dynamism for energy-efficient exascale computing), coordinated by TU-DRESDEN, reunites 7 European partners, started Sept. 1st and will run for 3 year

It aims at developing a software library to develop an integrated tool-suite and the READEX Programming Paradigm to exploit application domain knowledge, thus increasing significantly energy-efficiency of Exascale class computing systems. INTEL SAS France is a member of the READEX consortium and project participation is attached to the IPAG team of Intel present at the Exascale lab at Teratec.

HPC applications are usually highly compute intensive, however they also exhibit a large degree of dynamic behaviour, e.g., switching between communication phases and computation phases. Manually detecting and leveraging this dynamism to improve energy-efficiency is a tedious task that is commonly neglected by developers. Therefore, adjusting system –hardware and software- to application requirements is possible today but is seldom used in production by application developers However, using an automatic optimization approach, application dynamism can be detected at design-time and used to generate optimized system configurations. A light-weight run-time system will then detect this dynamic behaviour in production and switch parameter configurations to optimize performance and energy-efficiency of the application.

More information on www.readex.eu or Marie-christine.sawley@intel.com

December 16, 2015 - OpenHPC

The management of the HPC systems software stack is becoming more challenging with interdepending components and their respective pace of releases, from scheduling to monitoring agents, tools to scientific libraries, from diagnostics to corrective actions to name a few. Using OpenSource software building blocks has become more and more common, yielding more flexibility, allowing to deploy in house components based on local developments and to contain costs. Yet, the effort of assembling, checking compatibility and maintenance involve non trivial tasks that may impact productivity in the long term.

OpenHPC is a new, collaborative community effort that initiated from a desire to aggregate a number of common ingredients required to deploy and manage High Performance Computing (HPC) Linux clusters, including provisioning,  resource management, I/O clients, development tools, and a variety of scientific libraries.  The project was announced by the Linux Foundation on Nov. 12th with the download of the initial software stack. It is organized in away to share common solutions that we can build on as a community.  The community already encompasses a number of HPC software stakeholders, tool and OS developers, ISV, OEM, research labs, computing centres and is rapidly growing.

Intel, member of the OpenHPC community,  is committed to continue significant investments in this project while being open and welcoming to others who wish to contribute to this shared vision.  All OpenHPC packages will be tested by Intel on the latest Intel® Scalable System Framework (Intel® SSF) and will include the integration of Intel development and analysis tools (e.g., Intel Parallel Studio XE Cluster Edition).

More information can be found under

 

© Ter@tec - All rights reserved - Lawful mention