Phoenix, AZ. — SGI today announced at the SC03 conference its plans to deliver a revolutionary class of supercomputers that will catapult users into a new era of engineering achievement and scientific discovery for a large class of technical applications. Using a new concept SGI calls “multi-paradigm computing,” the groundbreaking machines will unite previously disparate computing architectures with SGI’s scalable shared-memory technology. Code-named “Project Ultraviolet,” SGI’s long-term vision is to improve productivity by creating the first supercomputers capable of supporting and combining different computational approaches, providing optimal performance specifically for technical applications, regardless of programming model or computational balance.
“SGI has reached the point where simply making a new computer faster than its predecessor, within the same architecture, is an insufficient goal. How powerful a computer really is depends on its applicability to a scientist’s or engineer’s particular problem,” said Dr. Eng Lim Goh, chief technology officer, SGI. “With the science-driven Project Ultraviolet and multi-paradigm computing, the primary goal is to make that user far more productive.”
Multi-paradigm computing serves to create a supercomputer capable of responding to the needs of a large class of technical applications. In pursuit of that goal, SGI has embarked on a system design that will not only incorporate elements of previously disparate computational architectures but also allow them to cooperate on the same data residing in scalable shared-memory.
“Today, scientists and engineers make machine purchase decisions that commit them to adapting or writing their applications, with life spans of 10 years or more, to match their chosen architecture. Moreover, the choice among clusters, scalar shared-memory, or vector architectures often restricts the types and sizes of data models that the adapted application can handle, due to inherent machine limitations,” said Goh. This is increasingly constraining progress, particularly with growing interests in the advanced computational techniques of stochastic, ensemble, multi-physics, multi-scale simulations and multi-disciplinary design optimization. Multi-paradigm computing will change that by efficiently supporting the different paradigms at a fundamental level. SGI will help scientists be more productive by letting them focus on science, not computer science.”
“Intel is excited to work with SGI on the Ultraviolet architecture,” said Justin Ratner, senior fellow, Intel. “Working through the Advanced Computing Program, Intel and SGI will develop new programming interfaces that are both more productive and highly scalable, enabling massively parallel rendering of large data models. The development of these technologies will make parallel processing simpler and easier for the Intel Itanium 2 market segment as well as for SGI’s high-end technical and creative users.”
In the coming months we will be announcing other collaborations, with government and industry, for Project Ultraviolet. Also SGI will unveil more specific plans for the next generation of SGI® systems that will take the first steps toward multi-paradigm computing, to be available in 2005. SGI plans the unveiling of the first true multi-paradigm computer in approximately 2007.
The pioneering implementation of multi-paradigm computing will grow from SGI’s expertise with scalable shared-memory systems. Since 1996, the SGI® Origin® family of servers and supercomputers has pushed the boundaries of system communication and management, allowing hundreds of microprocessors to work together in a common pool of data. Shared memory allows all of a computer’s resources to work on a single, large model simultaneously; without it, programmers need to break up models into smaller pieces, like a jigsaw puzzle, and then rely on complicated communications between the pieces to keep the entire model consistent as the data changes. Leading organizations such as the NASA Ames Research Center, Fleet Numerical Meteorological and Oceanographic Center, and the U.S. Department of Defense use Origin supercomputers because of their ability to analyze extremely large, complex scientific models with shared memory.
The SGI Altix family of servers combines industry-standard 64-bit Linux with the Intel® Itanium® 2 processor family and SGI® NUMAflex™ architecture to enable global shared memory systems from a few to hundreds of processors with up to 4 terabytes of shared memory, which is a first for Linux® OS-based computing. Powered by the third-generation NUMAflex supercomputing architecture, even the largest data sets can be handled and analyzed with ease and in record time for production workflows with the most demanding stability. The SGI® Altix™ 3000 family of servers is uniquely designed around this scalable shared-memory architecture that analyzes data sets as whole entities, without breaking them up into smaller segments to be handled by individual processors. The Altix architecture has proven ideal both for complex shared-memory applications running on a large single system image, and for communication-intensive applications optimized for clustering in throughput workflows.