High-Performance Technical Computing (known as HPTC) is the application of technology to highly complex scientific, engineering, and business analytics workloads. Traditionally, HPTC led to discoveries in the scientific community; however, it's rapidly making headway in the commercial market for its ability to help companies accelerate time-to-solution, design superior products, and reduce costs. Increasingly, companies are leveraging HPTC for a long list of applications, including business optimization, crash simulation, aircraft design, financial modeling, seismic analysis, climate modeling, and pharmaceutical development.1 Yesterday and Today
In the past, high-performance computing (HPC) was primarily the domain of supercomputers — dedicated, specialized number-crunching machines housed in a special environment. It was a monolithic system, complete with data storage, the compute power required to manipulate that data, and a way to extract the results of complex computations. In fact, only specially trained scientists were able to handle such a complex system. Yesterday's technical computing consisted of workstations used by engineers for production technical computing and by scientists for research. HPTC is essentially the marriage of HPC and technical computing. This union dramatically changes the way intensive computations are done by linking all the critical elements of data, compute, and graphics together. Today, HPTC employs an architecture comprised of several tiers that replaces the supercomputers and workstations of yesteryears and brings an abundance of compute power to mainstream commercial markets.
Broadly speaking, the current HPTC environment is comprised of three building blocks — massive data storage and retrieval, fast throughput computation, and high-end rendering of results through visualization techniques. Through a steady stream of technology innovations in networking products, chip design, routing algorithms, and bandwidth management, it's now possible to decouple each of the building blocks that comprise HPTC into an n-tiered architecture for scalability, availability, and cost-effective deployments. And thanks to innovations in grid computing, deploying this tiered architecture has gone beyond the client-server topology and is now a highly distributed network computing topology. Now HPTC applications are being powered by a host of heterogeneous resources of varied configurations — large and small — scattered throughout a network and pooled together through virtualization for maximum utilization.
Essentially, grid computing is at the heart of today's HPTC environment. It makes data, compute, and visualization resources available to HPTC applications — wherever and whenever they're needed. Grid computing also increases resource utilization, giving users maximum resources to power compute-intensive applications. What's more, the fusion of grid technology and portal computing technology means that the vast resources managed by the grid in HPTC environments can be made accessible from a desktop. This brings about an evolution in HPTC — today's users don't have to be highly specialized computer scientists to reap the benefits of HPTC. Through portal computing, HPTC is transformed from a tool used by the privileged few into a valuable compute resource that commercial users can utilize to solve today's most challenging business problems.
2 Grids Help HPTC Go Mainstream
Currently, we're in the midst of a data explosion where data is growing faster than processing power. In HPTC, data isn't just about having fatter pipelines; it's about data access and availability. In other words, it's about data movement. Enterprises are looking to clear the bottlenecks of hard-to-manage, isolated islands of storage. They need to find more effective ways to utilize storage assets. They want to share data without the latencies traditionally associated with common storage pools. The data grid provides an efficient and cost-effective solution by enabling the sharing of data and storage devices between heterogeneous computing platforms. Heterogeneous connectivity means providing access to independent compute platforms in a shared storage environment. This can result in significantly lower costs compared to using islands of storage that are hard-wired to each computational platform. The data grid provides the bandwidth as well as the performance for moving data, and very high efficiency for moving data over multiple pipes.
One of the biggest challenges IT faces is how to optimize data center efficiency and business productivity. When it comes to executing large computations, this is particularly critical. There is a great demand for systems that can manage a variety of compute infrastructures, so complex jobs can run in parallel to achieve faster results with greater efficiency and lower costs. This calls for increased use of throughput and parallel computing — running several applications at once across a cluster of multiprocessor and uniprocessor machines. The compute grid delivers an integrated hardware/software stack to simplify the process of deploying clusters for parallel computing. These integrated clusters can deliver hundreds of thousands of processors as a computing resource for HPTC applications.
No matter what the industry or how large the data, HPTC users need information to come to life. They need to visualize data, understand trends, and derive insights from information. A rich visual environment significantly advances problem-solving, prototyping, and research. Whether a user is modeling with large amounts of data, designing vehicles, or forecasting weather, the visualization grid provides solutions ranging from individual workstation-based graphics environments, to collaborative capabilities for workgroup visualization, to high-end visualization environments for large group viewing or collaboration. Such scalable, distributed visualization systems, combined with crisp, state-of-the-art graphics, enable designers, engineers, and researchers to interpret results from data more quickly, make better decisions, and reduce time-to-market.
