SPECrate2006: Alternatives Considered, Lessons Learned
- Cite this paper as:
- Henning J.L. (2009) SPECrate2006: Alternatives Considered, Lessons Learned. In: Kaeli D., Sachs K. (eds) Computer Performance Evaluation and Benchmarking. SBW 2009. Lecture Notes in Computer Science, vol 5419. Springer, Berlin, Heidelberg
Since 1992, SPEC has used multiple identical benchmarks to measure multi-processor performance. This “Homogeneous Capacity Method” (aka “SPECrate”) has been criticized on the grounds that real workloads are not homogeneous. Nevertheless, SPECrate provides a useful window into how systems perform when stressed by multiple requests for similar resources. This paper reviews SPECrate’s history, and several performance lessons learned using it: (1) a 4:1 performance gain for startup of a benchmark when I/O was reconfigured; (2) a benchmark that improved up to 2:1 when a TLB data structure was re-sized; and (3) a benchmark that improved by 52% after a change to NUMA page allocation. The SPEC CPU workloads usefully exposed several opportunities for performance improvement.
Unable to display preview. Download preview PDF.