Traditional DAQ vs DAQmx speed comparison

[Int_man]

I am currently using four (4) National PCI MIO 16E-1 boards in HP Workstation running Windows XP to simultaneously test four (4) medical devices at a time.  We are currently using Traditional DAQ and LabVIEW 8.2 to test our main line products.   Our test software is rigidly controlled and requires significant reesources for validation when changes are made.  We have purchased LabVEIW 2010 and it no longer supports traditional DAQ.  We know that sometime in the future we will have to make the transition to LabVIEW 2010 and DAQmx.

I need to perform a cost versus benefit analysis of converting the test system from LabVIEW 8.2 and Traditional DAQ to LabVIEW 2010 and DAQmx.  If the benefit is great enough I may be able to hire a consultant to expedite this process.

I know there are a lot of variables that factor into speed, and I know that I haven't provided many details but can anyone provide some ballpark estimates.  Is it 2%, 10%, 20%???  What is your best guess/estimate?

Thanks in Advance

[John_P1 · Accepted solution]

Hi Int_man,

Traditional DAQ 7.4.4 does support LabVIEW 2010, but to install Traditional DAQ to more than one version of LabVIEW at a time you will need to follow a procedure described in the readme.  Traditional DAQ does not  support newer OSes like Vista or Windows 7, and it does not support newer hardware like M or X Series.  Given your validation requirements it sounds like you might not be upgrading your OS or hardware in the near future, but if you ever decide to do so you would need to move to DAQmx.

Performance is a bit tricky...  You could see a perforamnce increase with an upgrade to DAQmx, but this depends completely upon your application.  I wouldn't feel comfortable providing any sort of estimate of performance increase without any information about the application itself.  For example, DAQmx provides for parallel threading which can often be used to increase application performance, but if your application spends the majority of time calling static digital writes in sequence you might see a slight performance decrease due to the overhead of a check that was implemented in DAQmx to prevent double-driving of lines.  Overall most of the functionality has improved in speed but it's not possible to say flat-out that "switching to DAQmx will give X% performance increase" for all cases.

Having said this, if you have to go through extensive validation anyway it might be a good time to reconsider the entire system as a whole (including DAQ hardware used).  Unless you're implementing feedback or require extensive user-input, the speed of the software driver calls could be made essentially irrelevant if you are are able to utilize hardware that allows you to buffer and trigger your necessary inputs and outputs.  As somebody who has been tasked to optimize sub-optimal systems (if that makes sense), I'd strongly encourage you to consider the entire application as a whole and where the most room for improvement lies.

Best Regards,