While high-performance computing systems that use application accelerators are still rare, they will be the norm rather than the exception in the near future. SAAHPC aims to bring together developers of computing accelerators and end-users of the technology to exchange ideas and learn about the latest developments in the field.
Reconfigurable systems have the potential to dramatically speed computing and spur scientific discovery. Someday it might be possible to build, on the fly, the machine that best fits the problem at hand. Instead of working with fixed "blocks" of computing power, scientists will have flexible "clay" that can be molded to perfectly fit their codes.
Until recently, computer performance improvements were largely driven by the ever-increasing clock speed of commodity microprocessors. As semiconductor devices shrink to enable higher clock rates, poor wire scaling becomes the performance-limiting bottleneck. In order to take advantage of the shrinking transistor size and alleviate the poor wire scaling, semiconductor manufacturers resort to placing multiple copies of the same design on a single die instead of increasing the clock speed. Multi-core chips, unheard of just a few years ago, are a standard today. While the general-purpose processors are still restricted to just a few cores per die, more specialized chips, such as stream processors used in computer graphics applications, contain hundreds of smaller well-tuned SIMD-enabled cores capable of running thousands of execution threads.