However, the computing future presents new challenges. As component sizes continue to shrink and the processing speed of central processing units (CPUs) has plateaued, CPU vendors are adding additional processing units or “cores” to a single chip (a form of parallelism) to achieve performance gains that will present significant challenges to the development of effective and efficient scalable software on a node. Next-generation supercomputers will rely on many thousands of multicore nodes working together, presenting numerous challenges in the areas of resiliency, power, cooling, and scalability.

Multicore processors and accelerators, including inexpensive and abundant graphical processing units, are changing the landscape of computing through a new era of on-chip parallelism. Also, programming models, libraries and compilers that automate parallelization and threading of sequential code, and a new generation of computer programmers who think and code in a multithreaded parallel way, would open these future high-performance computing (HPC) platforms to a broad set of application developers.

Once the purview of technologically sophisticated and wealthy nations, modeling and simulation capabilities are now well within the reach of most state and nonstate actors. Advances in computation, representational algorithms, and the software to harness these advances will continue and will be accessible to all developed and developing societies. However, the increased ubiquity of computing power has lowered the threshold for obtaining effective modeling and simulation capabilities. No state or nonstate actor is likely to maintain a sustainable, strategic, comparative advantage in access to these capabilities. From a national security and an HPC leadership perspective, there will be international competition to develop the capacity for usable exascale2 computing, with specific “breakthrough” points to be achieved.

The scientific and systems skills needed to create algorithms that capture the underlying scientific and system phenomena to be represented by models and simulations are important elements of capability. There is a notable lack of highly skilled computational scientists and engineers able to fully leverage the current state of the art in HPC for science-based modeling and simulation (NRC, 2007; WTEC, 2009). New computing architectures for games will continue to lead to new capabilities provided by modeling and simulation, including increased predictive accuracy. As computers have become faster, models have increased in fidelity, causing games to become more realistic and accurate. Second, the increasing fidelity of games has led game makers to seek the manufacture of higher-performance computer chips. As such, the need for a skilled workforce in modeling and simulation, and in simulation-based engineering and science generally, will become increasingly important to national security to take advantage of improvements in computing speed and accuracy.3

GAMES: BEYOND ENTERTAINMENT

The games industry has evolved over time from purely commercial entertainment to more recent applications of games by civil, commercial, and military organizations (WTEC, 2009). The games industry and the personal computers industry have provided innovations in graphics and PC hardware that have led, for example, to PC-based simulations. Until relatively recently, most military and government simulations were performed on workstations (i.e., SGI, Sun) and on specially designed hardware (i.e., Evans & Sutherland, Delta Graphics).

Evidence indicates that the use of games might have substantial impacts on human and group behav-

2

Systems that can handle a million trillion, or 1018, floating-point calculations per second.

3

The committee notes that a good reference for simulation-based engineering is the National Science Foundation report Simulation-Based Engineering Science, a summary of which is provided online at http://www.nsf.gov/attachments/106803/public/TO_SBES_Debrief_050306.pdf. Last accessed on October 14, 2009.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement