REDMOND, Wash. – Oct. 13, 2009 – Computing in the 21st century is experiencing a profound transformation. The increasing use of multicore chips and the move to huge facilities hosting data in the cloud have created an environment in which bold new approaches are needed, say thought leaders at Microsoft.
Dan Reed, corporate vice president of the eXtreme Computing Group.
Enter Microsoft's new eXtreme Computing Group (XCG).
Formed in June with the goal of developing radical new approaches to ultra-scale and high-performance computing hardware and software, the group's research activities span a variety of cutting-edge computing efforts: security, cryptography, operating-system design, parallel-programming models, cloud software, data center architectures, specialty hardware accelerators, and quantum computing.
By unifying research and incubation efforts, the group will help develop radical new approaches to building hardware and software, says Dan Reed, corporate vice president of XCG. In essence, it's an "intentional R&D group" that sits between basic research and product development, funneling ideas from Microsoft Research into product teams across the company, he said. In addition to contributing code and building prototypes, Reed and his team of researchers, engineers, and developers will pursue outcome-driven research and development.
"Our objective is to look at strategic needs and opportunities that cut across product groups and find technology solutions to those problems," Reed says.
Gustav is an eXtreme Computing Group project aimed at producing painting software that is easy to use, leverages the multiple cores of future computers, and provides a natural, immersive user experience.
XCG will tackle challenges such as cryptography and parallel-programming models with rapid, large-scale prototyping and testing. That testing will help it transfer new technologies to Microsoft partners and product teams. "It's not just 'let's look at this problem and figure out new alternatives,'" Reed says. "It's 'look at the problem, figure out some new alternatives, build some prototypes of those alternatives, validate them, and then push them into production.'"
As the name suggests, "extreme computing" focuses on moving beyond current limits of computing and reshaping fundamental assumptions and practices, Reed says. The team's charter also includes engaging academia and government agencies on the use of Microsoft technologies for research and production.
Multicore processing alone promises to revolutionize computing. A multicore processor includes two or more "cores," or processors, on a single chip. That means systems can perform computations simultaneously rather than sequentially. As parallel computing becomes widespread the implications will be far reaching. "We have an incredible opportunity, one that comes along once every two or three generations, to reinvent computing," Reed says. "The opportunity for Microsoft is to leverage those technologies to do that."
Reed predicts that there will be a huge set of technology changes on the hardware level—and even more to systems software and next-generation applications. Multicore issues will affect Windows and other business products, and XCG is helping to prepare those product teams to cope with the new reality.
Earlier this year, the eXtreme Computing Group demonstrated a hardware/software prototype based on Intel’s low-power Atom processors, which are starting to appear in netbooks.
The group operates on a five- to seven-year time horizon, and, along the way, it will spin off technologies that have shorter-term significance, Reed says. For example, earlier this year, the team demonstrated a hardware/software prototype based on Intel's low-power Atom processors, which are starting to appear in netbooks. An intelligent energy-management system could turn processors on and off automatically while still delivering performance, and XCG is working with the Windows Azure team to transfer the energy-management software it has developed.
Reed joined Microsoft in December 2007 as a strategist in scalable and multicore computing. In February 2008, he assumed the responsibility of directing the new Cloud Computing Futures Initiative, which explores new approaches to cloud services and data center design, including ways to reduce hardware costs and power consumption and to increase data centers' adaptability and resilience to failure.
Before joining Microsoft, Reed was Gutgsell professor and head of the Department of Computer Science and director of the National Center for Supercomputing Applications at the University of Illinois. He also was Chancellor's Eminent Professor at the University of North Carolina at Chapel Hill, founding director of the university's Renaissance Computing Institute, and the university's chancellor's senior advisor for strategy and innovation.
Reed said he was able to make an easy transition from academia into the private sector, though people often ask why he wanted to leave academia after 25 years to move into the corporate world. For him, the answer was simple.
"This is the only place on the planet where one can tackle these problems at this scale," Reed says. "The whole goal for the XCG is to position Microsoft to do things that no one else on the planet can do."