Archive for January, 2014

The Dangers of Benchmarks

January 17, 2014
  • Benchmarks are not a true reflection of application performance.
  • Some benchmarks are closer to real world performance than others. Other benchmarks can generate very artificial workloads and which may lead you to incorrect conclusions.
  • Use benchmarks as one data point for performance comparisons between systems or troubleshooting performance issues.
  • As always, the more data points you have, the more evidence you have to support a particular conclusion.