The speed of currents at the ocean surface simulated with a three- dimensional global ocean model developed for the Thinking Machines CM-2 and CM-5 at Los Alamos National Laboratory. High speeds are in red, low in blue. Eddies are evident in the Gulf Stream along the east coast of North America and in the tropics. The spatial resolution of the computer model is 0.5 degrees in latitude and longitude with 20 vertical levels; realistic ocean bottom topography is used. The work was jointly sponsored by the HPCC and the DOE CHAMMP (Computer Hardware, Advanced Mathematics, and Model Physics) Program.
This section presents examples of on-going Grand Challenge applications research and emerging HPCC hardware and software technologies being developed in support of real-world problems and the HPCC Program.
Only within the past several decades have advances in high performance computing and networking technologies merged with long standing theoretical and experimental methods to yield a powerful new computational approach to scientific inquiry. This approach has enabled scientists and engineers to transcend many of the limitations inherent in the more traditional practices. The demonstrated success of computational science in a wide variety of problem areas has led to an infusion of high performance computing technologies and techniques into mainstream scientific practice. An increasing number of researchers, representing almost every discipline, continues to probe new and complex problem areas -- many of pressing societal concern.
The computational approach adds another dimension to research methods by allowing the researcher to create a mathematical model of some aspect of reality. Solving the model entails translating its equations into a form capable of being programmed and executed on a high performance computing system. Algorithms that structure input data and specify the manner in which the calculations are to be performed are the primary building blocks of the computational model. By exercising the model over broad data ranges and parameter spaces, a picture -- albeit a simulated one -- of the real phenomena emerges. To the extent that it is complete and accurate, this picture is useful in describing reality as it exists and perhaps more importantly, predicting change. In many cases such as those described in this section, scientists are able to reach and extend the understanding of phenomena far beyond what is possible through pure reasoning and observation.
The success of this approach directly depends upon computing and data management capability. Input data ranges and parameter spaces may be mathematical continua -- infinite in dimension. It may require billions of calculations to produce a single solution point, and some phenomena require billions of solution points to approach a useful level of completeness and accuracy. Thus the computational requirements of many applications are potentially boundless. Similarly, the "answers" associated with a simulation may comprise terabytes of data. Scientific visualization, with its ability to store and interpret data, has come to play an indispensable role in the overall success of computational research.
The strengths and limitations of the computational approach are readily evident: a simulation can represent reality only to the degree that it both holds to the physical laws of nature and captures the inherent complexity and detail of that which it attempts to represent. However, for many problems this is the only approach available. Scientific instruments for observation and measurement confront limits -- either those imposed by the nature or location of the object of interest, or by the economics of producing the instrument or conducting the experiment. There are few alternatives to the computational approach for studying aspects of nature lying at the extremes of measurability -- those that are very small, very large, very fast, very slow, very close, very far, and so on. Yet the universe is filled with such phenomena that affect our daily lives.
The examples presented here are only a small subset of the early achievements to come out of the High Performance Computing and Communications Program -- a hint of the promise of HPCC methods and technologies applied to long standing problems of science and engineering -- and humanity.