High-performance computing (HPC) has moved from the domain of government and academic laboratories to being an essential component of the design process. Today, it is almost unthinkable to develop the key components of a car, airplane or even many consumer products without computer-assisted structural or impact analysis.
Much of the credit for this can be given to the introduction, about six years ago, of high-performance, open-source Linux clusters in place of proprietary UNIX systems. These Linux supercomputers delivered dramatically lower costs, thereby enabling a wider market to take advantage of high-performance computing.
With the introduction of Windows Compute Cluster Server, it is natural for many to question whether the tremendous price advantage of Linux and open systems still outweighs all other considerations. The truth is while UNIX still has a fairly healthy legacy install base and Windows provides some advantages at a very entry level, Linux remains not only the most viable solution for most design engineers, but is actually more attractive than ever.
There are several reasons for this: First, due to the quick-paced nature of open-source, Linux-based systems offer the most diverse choices in hardware and software components. Standardization of these components drives higher price-performance and improves already high system reliability. Second, selecting and purchasing an HPC system is a much more delicate proposition than purchasing a desktop computer. Small variations in system architecture can have a disproportionately large impact on time-to-production, system throughput and overall price-performance. There are several characteristics that make high-performance Linux clusters ideally suited to overcome this problem. To begin with, the diversity of hardware and software enables system architects to more easily define a system optimized for specific tasks or application codes. With years of experience under their belts, vendors and architects of high-performance Linux clusters are better equipped than ever to design stable, tuned systems that deliver the desired price-performance and enable customers to get the most out of their application licenses.
Finally, Linux-based systems have matured to the point that vendors are beginning to deliver added innovation and functionality. The first generation of Linux clusters often arrived at the customer site as a generic compute platform with no specific tuning or application focus — and often with minimal integration and testing for even baseline system stability. Simply putting a system into production could take months. Now, however, vendors are beginning to ship actual production-ready Linux “systems” as opposed to Linux “Pile-of-PC” clusters. This allows design engineers to get the price-performance of Linux coupled with the simplicity of management, reliability, availability and production-ready characteristics of proprietary UNIX systems.
So where does Windows fit in?
Windows has a lot to offer design engineers, especially those with more limited resources or goals. UNIX still has a lot to offer for many legacy HPC applications. Both Windows and UNIX require more work to be able to deliver the same functionality and compelling price-performance value of Linux. The HPC market is more open and competitive than it has ever been, but for design engineers focusing on solving their design challenges, it's clear that Linux is still the best choice for today and the foreseeable future.