An exascale workload study

Amdahl's law has been one of the factors influencing speedup in high performance computing over the last few decades. While Amdahl's approach of optimizing (10% of the code is where 90% of the execution time is spent) has worked very well in the past, new challenges related to emerging exascale heterogeneous architectures, combined with stringent power and energy limitations, require a new architectural paradigm. The 10x10 approach is an effort in this direction. The 90-10 approach viewed application sets as monolithic entities with common cases and focused on applying broad architectural improvements impacting those common cases. In the 10x10 paradigm, the objective is to identify the top ten distinct dominant characteristics in a set of applications. One could then build and exploit customized architectures (accelerators) and tools, best suited to optimize each dominant characteristic in the application domain. Every application will typically be composed of multiple characteristics and thus will use several of the customized accelerators/tools during its various execution phases. The goal is to ensure that the application runs efficiently and that the architecture is used in an energy-efficient manner. In this poster, we describe our initial steps and methodologies for defining and actualizing the 10x10 approach.