TC3 - Certification - TF Benchmarking
PLCopen TC3 - TF Benchmarking – released for comments
As of December 22, 2005, the technical specification of this PLCopen Task Force has been released to the community for comments till June 30, 2006. With this release we want feedback form a broader community about its usability for their environments before developing the applicable test software.
The published version is 0.4, and is combined with a feedback form.
The released document is available at the PLCopen website under TC3 – Certification / Task Force Benchmarking for free-of-charge download. Please use the included form for your feedback, to improve the usability of this for you.
A benchmark is a reproducible, portable test to measure the performance of a given system in comparison to other systems. For PLC systems there are no defined benchmarks. The only common measurement for PLC performance is the execution speed of 1000 Boolean operations. This number is no good benchmark, because one cannot derive the performance of the PLC in a typical application from it, and as such it is not comparable between systems due to the lack of the definition of the test conditions.
There are two main objectives, when you use a benchmark.
1. To estimate the performance of the PLC in your own application
2. To compare the performance of the PLC with other PLCs and find out the specific strength and weakness of a given system
The paper defines two different sets of benchmarks to meet the objectives described above.
The first set of benchmarks defines five different types of applications, which are typical for the usage of a PLC. Most applications belong either to one of these types or consist of a mixture of them.
The second set of benchmarks measures each language feature of the IEC 61131-3 separately. The goal of this test is to exclude effects of other features to the most possible extent.
|
