Execution time of IVI Steps

[Olivier_Capiaux]

Hello,

We have made some benchmarks to decide the usefulness of TestStand IVI steps versus Labview ones with a PXI 4070 ( DMM ).
The Labview steps ( using also IVI access !!! ) are 30% faster than direct TestStand IVI steps.
May someone help us to understand such difference ?

Thanks a lot.

[Scott_Richardson]

Oliver -
The TestStand step types have additional overhead because all the configuration information stored on the TestStand step must be evaluated and passed to the a VB component.  Because the TestStand execution is multi-thread apartment and the VB component must be single thread apartment, some marshalling of data is necessary via COM.  Lastly the VB component is not optimized as the IVI VIs that you are using.  The TestStand IVI steps are not intended for speed, more easy of use. Hope this helps. 

[Olivier_Capiaux]

Thank you Scott.

    In fact, we have seen that the "niDmm_32.dll" was called in Labview ( so not an IVI dll ) and the "iviDmm.dll" in TestStand". That should explain this big difference.

    So I would like to know, from your expert point of view, which is  the best way to have the fastest access to PXI DAQ and Switches Ressources from TestStand to reduce the test times.

Thanks a lot

Regards

Olivier

[Scott_Richardson]

Olivier -
I do not know alot about DAQ and switching speed issues for software. I would assume that controlling the hardware directly from a development environment, like LabVIEW, would be optimal, but I think these questions are better sent to their respective forums to get thorough responses for each domain.