Cross-profiling for Java processors
Software—Practice & Experience
| Hi-index | 0.00 |
Most processors are used in embedded systems, where the processorarchitectures are diverse due to optimizations for differentapplication domains. The main challenge for embedded systemprocessors is the right balance between performance and chip size,which directly relates to cost. An early estimation of theperformance for a new design is of paramount importance. In thispaper we propose cross-profiling for that performance estimation,which can be accomplished very early in the design phase. We evaluateour approach in the context of a Java processor for embedded systemsusing cross-profiling on a standard desktop Java virtual machine. Weexplore the performance impact of various processor design choicesand optimizations, such as different caches strategies or pipelineorganizations, and come up with an improved processor design thatyields speedups of up to 40\% on standard Java benchmarks. Comparingthe generated cross-profiles with the execution of benchmarks in realhardware confirms that our approach is sound.