A Performance Chart is a tool for comparing different servers
It allows you to know the improvement you can expect from an upgrade of your machine. Each machine is rated on performance criteria such as:
- floating point calculations,
- web server performance,
- client/server interaction, or
- raw computing power.
Results are then displayed in an easy to interpret format. A rating is created using a model, simulating a computing environment.
IBM use a model where a machine is selected as the base, and other machines are benchmarked against it and given a comparative rating. Some comparisons are specific to particular types of machines.
- IBM uses ROLTP to measure performance of the pSeries and RS/6000,
- rPerf to measure the later pSeries and RS/6000, and
- CPW to compare iSeries/AS400.
Some independent organizations such as TPC and SPEC also offer a range of comparisons across different manufacturers.
Transactions per minute of cost/performance ratios are examples of other common industry benchmarks.
- ROLTP – Relative On Line Transaction Performance
- Uses the 7011/250 as the base model to compare commercial processing power.
- rPerf – Relative Performance
- This has replaced ROLTP as a measure of commercial processing power, using a more modern model. Uses the 7026/B80/p640 as the base model.
- CPW – Commercial Processing Workload
- CPW replaces the earlier Relative Performance, which stemmed from 9404 B10 as one of the earliest AS/400′s
- SPEC – Standard Performance Evaluation Corporation
- An organization founded in 1988 to establish standard benchmarks for computers. SPEC has various measures of performance.
Next time you are talking to us at Computer Merchants, ask us for a copy of some ROLTP, rPerf or CPW charts.
See Also: Comparing IBM System i And System p