This is the third article in a series of articles on Linux Benchmarking, to be published by the Linux Gazette. The first two articles showed how to successfully run synthetic or application benchmarks to produce accurate, significant and relevant data. The present article deals with the correct interpretation of this data, and it also presents nbench-byte 2.1, a modern CPU benchmark suite.

1. Contributors

2. Benchmarking vs. benchmarketing

• 2.1 The scientific/quantitative approach
• 2.2 The benchmarketing approach

3. Benchmarks for SMP systems

4. GNU/Linux specifics

5. An example of correct/incorrect interpretation of results

6. In the next article

7. Notes