El día 10 de junio de 2008 8:41, Rafa Grimán <rafagriman@gmail.com> escribió:
Hola :)
El Tuesday 10 June 2008, Carlos E. R. escribió:
El 2008-06-10 a las 11:26 +0200, Rafa Grimán escribió:
No es por quitarles mérito ... pero no es la primera supercomputadora híbrida ... Según la web de IBM, pone:
¿A que se refieren con lo de hibrida? No me he dado cuenta de si se ha dicho :-?
La computación híbrida es cuando mezclas procesadores de distinto tipo. En el caso de RoadRunner:
"Two IBM QS22 blade servers and one IBM LS21 blade server are combined into a specialized "tri-blade" configuration for Roadrunner."
Es decir:
- RoadRunner es un cluster, no es _una_ máquina - es un cluster compuesto por tri-blades (aquí es donde entra la computación híbrida) - en el triblade se mezclan procesadores Cell y AMD
No puedo dar muchos más detalles por que no sé cómo se interconecta todo, la memoria, los buses, ...
Como decía esto no es nuevo ni es la primera vez que se hace,
Error! Hay algo nuevo, y es la plataforma de interconección Torrenza: http://www.idg.es/computerworld/articulo.asp?id=179042 AMD anunció para su próxima generación de procesadores Opteron la incorporación de 3 nuevas tecnologías como son Trinity, Radien y Torrenza. Pues bien, hoy, ya se conocen los pormenores de ésta última, que está disponible en todo el mundo y que la propia AMD definide como una plataforma multisocket abierta, destinada al entorno servidor, plataforma en la cual otras compañías importantes pueden también montar sus propios procesadores o co-procesadores multipropósito, usando para éste propósito, tecnologías de AMD como AMD64, Direct Connect y HyperTransport. Dentro de la propia AMD, nadie duda de que la nueva plataforma tendrá un efecto positivo en la labor de los administradores de centros de datos ya que, según señala Marty Seyer, vicepresidente senior del departamento comercial de AMD, "se verán beneficiados por la cooperación optimizada a nivel de plataforma, con nuevos niveles de estabilidad, posibilidad de ampliación, flexibilidad y propiedades de plataforma para la infraestructura de su servidor, entre otras características". Aunque en teoría Torrenza "está orientada a fabricantes de CPUs, tarjetas gráficas o chipsets", como señala Alejandro Ruiz Bueno, director de desarrollo de negocio de AMD para España y Portugal, son ya varios los proveedores de la industria como "HP, IBM, Dell, Sun Microsystems, Fujitsu Siemens Computer, Tyan, o Supermicro, entre otros, los que han mostrado su interés por trabajar sobre Torrenza como plataforma abierta", tras haber aplaudido enérgicamente la iniciativa de AMD. Primera referencia AMD también ha hecho pública la primera referencia en la utilización de esta tecnología, que no será otra que el superordenador Roadrunner de IBM, desarrollado por el Gigante Azul para el departamento de Administración de la Seguridad Nacional de Energía Nuclear en Estados Unidos. Este superordenador llevará 16 mil procesadores AMD Opteron en servidores x3755 con más de 16 mil procesadores Cell de IBM en unidades Blade Center H con enlaces Hypertransport entre ellos. El Roadrunner confiará en los procesadores Opteron de AMD para procesar los datos de entrada y salida (IO) y la actividad de las comunicaciones mientras que los cálculos más complejos los harán los procesadores Cell de IBM, que cuentan con 8 cores cada uno. Aunque el fabricante no lanza previsiones en cuanto a la adopción de esta plataforma, el portavoz local insiste en que "desde AMD España ayudaremos a aquellos que quieran empezar a desarrollar sobre Torrenza para tener referencias lo antes posible", augurando un buen futuro a este lanzamiento. http://www-03.ibm.com/press/us/en/pressrelease/20210.wss How it Works Roadrunner's hybrid design will allow the system to segment complex mathematical equations, routing each segment to the part of the system that can most efficiently handle it. Typical compute processes, file IO, and communication activity will be handled by AMD Opteron processors while more complex and repetitive elements -- ones that traditionally consume the majority of supercomputer resources -- will be directed to the more than 16,000 Cell B.E. processors. Designed originally for gaming platforms, where intense graphics and real-time responsiveness are key, the Cell B.E. processor is ideal to speed Roadrunner through intense mathematical problems. "This new supercomputer demonstrates a commitment to achieve a major advance in technological capability that will help enable scientists and businesses solve the most challenging problems," said Bill Zeitler, senior vice president, IBM Systems and Technology Group. "Los Alamos is a valued partner as we embark on this exciting journey." "This installation with Los Alamos and IBM demonstrates the compelling benefits from industry leaders innovating around an open platform; in this case IBM and AMD collaborating in the use of AMD Opteron and the Cell B.E. processor to build powerful systems for highly specific Los Alamos Labs workloads," said Marty Seyer, senior vice president, Commercial Segment, AMD. "This is an excellent demonstration of Torrenza in action -- building on the performance and performance-per-watt advantages AMD delivers to create incredible value in leveraging HyperTransport technology to redefine how different systems, based on different processor platforms, can communicate with each other to solve some of the most complex computing problems." http://enterprise.amd.com/us-en/AMD-Business/Technology-Home/Torrenza.aspx Torrenza represents the industry's first open, customer-centric x86 innovation platform, capitalizing on the Direct Connect Architecture and the AMD64 platform to enable other processor and hardware providers to innovate within a common ecosystem. This collaboration allows partners with acceleration technology to increase system performance, scalability, and flexibility. And because end customers value openness and freedom of choice, Torrenza helps enrich the IT industry. At AMD, we have made a commitment to open AMD64 technology and the AMD Opteron™ processor platform to collaboration from a broad community of innovators, including hardware manufacturers and software developers. As an open environment, Torrenza was conceived to enable: * Tight coupling of accelerator coprocessors and option cards to AMD64 technology-based systems * Enhanced protocols to make accelerators more effective AMD encourages innovation by the sharing of technology. As an example, the Torrenza Innovation Socket enables OEMs to develop their own silicon to take full advantage of an x86 environment and the accompanying economics associated with packaging, chipsets, and motherboards designs. Through the Torrenza initiative, AMD has shared the AMD64 processor socket design and specifications to OEMs to allow collaboration on specifications and to take full advantage of Direct Connect Architecture. Torrenza is designed to reduce complexity for customers while increasing the pace of innovation in both silicon and platforms. Collaboration between AMD and the community of innovators will help provide datacenter managers with new levels of platform stability, upgradeability, and flexibility for their server infrastructure. Stream Computing Stream computing harnesses the tremendous processing power of graphic processors for high performance, data-intensive computing in a wide range of scientific, business, and consumer applications. This provides organizations with the ability to process massive amounts of information in significantly less time. * Learn more about stream computing. Innovation Community Leading server OEMs that develop silicon, including Cray, IBM, and Sun, have endorsed Torrenza as an open innovation initiative. Discover how AMD is creating opportunities for the global innovation community. Advanced Research AMD has teamed with the Computer Architecture Group at the University of Mannheim to establish the Center of Excellence for HyperTransport. Research at the Center will benefit the academic community and the next generation of solutions that leverages HyperTransport™ technology. The Center of Excellence has released an HTX reference board for universities and companies that research compute-intensive testing and design applications. Research projects focus on both non-coherent and coherent HyperTransport, helping to support an open standards approach to innovation. The University of Mannheim is also a licensee of coherent HyperTransport. HyperTransport™ Technology Consortium The HyperTransport Technology Consortium controls, licenses, and promotes HyperTransport technology industry-wide. Since its inception in 2001, the Consortium has expanded to more than 60 members worldwide, including technology leaders such as AMD, Apple, Broadcomm, Cisco, Cray, Dell, IBM, NEC, Nvidia, Renesas, and Sun Microsystems. OpenFPGA Consortium OpenFPGA is a non-profit consortium focused on accelerating the adoption and incorporation of reconfigurable computing solutions in high performance and enterprise computing. The consortium serves the worldwide community through collaborative development of the best practices and industry standards in reconfigurable computing. More than 200 organizations and 40 countries are represented through various working groups of OpenFPGA. White Papers Find out how acceleration technologies give you a competitive advantage: * Protocol enhancements for Tightly Coupled Accelerators (PDF 230KB) * AMD's holistic vision for the future: Torrenza enables platform innovation (PDF 592KB) * Acceleration technologies need to achieve cost-effective supercomputing performance for advanced applications (PDF 440KB) --------------------------------------------------------------------- Para dar de baja la suscripción, mande un mensaje a: opensuse-es+unsubscribe@opensuse.org Para obtener el resto de direcciones-comando, mande un mensaje a: opensuse-es+help@opensuse.org