Englobe Technologies & Systems
ENGLOBE Technologies and Systems nace como consecuencia de las nuevas demandas que se están generando en el nuevo mercado de la ingeniería.
ScaleMP es líder en virtualización para computación de alta gama, proporcionando mayor rendimiento con un menor coste total de propiedad. La innovadora arquitectura Versatile SMP (vSMP) agrupa múltiples sistemas x86 en un único sistema x86 virtual, poniendo a su disposición un sistema multiprocesador simétrico (SMP) estándar de alta gama. Utilizando software para reemplazar el hardware y componentes a medida, ScaleMP ofrece un nuevo y revolucionario paradigma de computación.
vSMP Foundation™ Standalone, optimizado para el Cray CX1, permite la creación de sistemas x86 de memoria compartida con hasta 32 zócalos mediante la agrupación de los procesadores, la memoria y los dispositivos E/S de sistemas independientes de 2 zócalos, inteconectados por medio de InfiniBand. El sistema de resultante, con hasta 128 núcleos de procesador y hasta 1TB de RAM, ofrece inigualables ventajas en cuanto a precio y rendimiento para el mercado de la supercomputación (HPC).
Crear de forma transparente un sistema de memoria compartida de 4 procesadores interconectando dos sistemas de 2 procesadores
Energy companies and energy services companies are always looking for ways to reduce risk in seeking new sources of energy and improve operational efficiencies.
In addition to the challenges of exploring, extracting, refining and delivering energy sources they are also grappling with fluctuating margins, challenging political environments, regulatory constraints and capacity pressures. To succeed in such a difficult environment, Energy companies have turned to solutions based on vSMP Foundation across the entire workflow: to increase production yields, discover new energy sources, and run simulations. As with other industries, solutions based on vSMP Foundation have the advantage of offering high memory and memory bandwidth required by most reservoir and volume interpretation applications and the higher number of processors required by the seismic processing applications. In addition, leveraging the power of commodity x86 infrastructure coupled with the significantly lower costs of system hardware provides the high performance they expect and the significantly lower costs Energy companies appreciate.
Energy companies and energy services companies are always looking for ways to reduce risk in seeking new sources of energy and improve operational efficiencies.
In addition to the challenges of exploring, extracting, refining and delivering energy sources they are also grappling with fluctuating margins, challenging political environments, regulatory constraints and capacity pressures. To succeed in such a difficult environment, Energy companies have turned to solutions based on vSMP Foundation across the entire workflow: to increase production yields, discover new energy sources, and run simulations. As with other industries, solutions based on vSMP Foundation have the advantage of offering high memory and memory bandwidth required by most reservoir and volume interpretation applications and the higher number of processors required by the seismic processing applications. In addition, leveraging the power of commodity x86 infrastructure coupled with the significantly lower costs of system hardware provides the high performance they expect and the significantly lower costs Energy companies appreciate.
Computational chemistry is part of a disciplinary approach to predict the nature and function of new chemical compounds. By creating new molecular structures, scientists can create new products and improve existing products in a cost effective manner. This has a profound impact on the development and enhancement of products in a variety of industries including pharmaceuticals, plastics, glass, metal, paint and manufacturing process including aerospace and automobiles among others.
Computational chemistry shortens the development cycle for new drugs, and can save millions of dollars through early-stage simulations. Significant improvements in computer hardware and software allow researchers to perform highly complex analysis; predicting the properties of new chemical compounds and materials before any laboratory effort.
Computational structural and chemical analysis is not new and many applications have been developed over the years. A significant characteristic of this type of work is that scientists use a variety of different applications, including commercial, in-house custom, legacy and parallel applications using OpenMP, PVM, MPI and other types of message passing architectures. As such, computational chemists require well-balanced systems that are flexible and can run different application types. Solutions based on vSMP Foundation are ideal for these environments. The shared memory architecture can run all these applications with good performance, sometime leveraging the large compute, memory, or bandwidth configurations or a combination of each.
Solutions based on vSMP Foundation provide high performance coupled with lower management costs. They are particularly well suited in environments where computational chemists do not have dedicated IT staff or need to publish results fast for new innovative applications.
To accelerate design and product development, and reduce time-to-market, manufacturing organizations are increasingly relying on IT to deliver infrastructure for simulations with increased model sophistication and detail. These simulations drive the need for faster compute that traditional dual- and quad-processor systems, with their limited memory, can no longer meet. To overcome these system limitations, software vendors have increasingly invested in scaling their applications to take advantage of the largest CPU solutions. These can be run either on shared memory processing (SMP) systems or on distributed memory systems (clusters). Market analysis shows that, when taking into account application scalability and software-licensing models, optimum system configurations are in the range of 4- to 16-processor and 128 GB to 512 GB. This makes solutions based on ScaleMP’s vSMP Foundation aggregation platform an ideal choice for these environments.
Scientists and Engineers using numerical simulations applications such as MATLAB, R or Mathematica for research in signal processing, financial engineering, mathematical modeling, manufacturing and more, can greatly benefit from solutions based on vSMP Foundation. Most vendors provide both single processor and multi-processor versions of their applications, and as such, users can run either single processes using all the memory in the system, multiple processes in parallel that share the memory, or one or more jobs running in multi-processors mode. This flexibility allows engineers and researchers to run more simulations with greater flexibility, without the risk of running out of memory for large jobs, or being able to run demanding jobs on the entire system using parallel versions of the applications when available. This dramatically accelerates their application/simulation development efforts, for faster time to insight and products.
