The HPCC benchmark suite was initially developed for the DARPA's HPCS program 1 to provide a set of standardized hardware probes based on commonly occurring computational software kernels. The HPCS program has initiated a fundamental reassessment of how we define and measure performance, programmability, portability, robustness and, ultimately, productivity in the high-end domain. Consequently, the suite was aimed to both provide conceptual expression of the underlying computation as well as be applicable to a broad spectrum of computational science fields. Clearly, a number of compromises must have lead to the current form of the suite given such a broad scope of design requirements. HPCC was designed to approximately bound computations of high and low spatial and temporal locality (see Figure 1 which gives the conceptual design space for the HPCC component tests). In addition, because the HPCC tests consist of simple mathematical operations, this provides a unique opportunity to look at language and parallel programming model issues. As such, the benchmark is to serve both the system user and designer communities 2.
Finally, Figure 2 shows a generic memory subsystem and how each level of the hierarchy is tested by the HPCC software and what are the design goals of the future HPCS system - these are the projected target performance numbers that are to come out of the wining HPCS vendor designs.
|
|