Next: Summary of Chapter 1
Up: Performance Metrics
Previous: ...
Use a set of several real applications, covering a range of
application areas and program structure
- Applications are still not average programs, so results can still not reliably be
extrapolated.
- Misleading to combine performance on different programs since
unclear how they should be weighed.
- Compiler technology can still make big difference
- Still fails to account for effort needed in manual performance tuning for
many architectures
- Eg SPEC
www.specbench.org (rate vs time)
- NAS parallel b/marks
www.nas.nasa.gov/Software/NPB
- MediaBench
www.cs.ucla.edu/~leec/mediabench