Performances of a release application are different with or without the LabWindows/CVI IDE

[Benjamin_C]

Hi,

I am noticing a strange behavior with a LabWindows/CVI application that I have just developed. I would like your opinion about this.

First of all, I use LabWindows/CVI 8.1 with Windows XP. My application is an executable that calls some CVI DLLs. The application is basically a multithread application that use DAQmx. My application is supposed to perform operations within a defined time. The execution speed is an important factor for me.

Here is the issue...
If I launch the realease executable only, the operations are performed in about 15 seconds. If I launch the CVI IDE then the executable (outside the CVI IDE), the operations are performed in 12 seconds. The execution speed is faster if the LabWindows/CVI IDE is running!

Is the runtime somehow affected by the CVI IDE? I don't know.

I am still investigating this issue. I don't expect any of you to give me a direct and clear explanation of this issue with so few data. But any clue or idea of what would cause such a behavior would be a great help!

Thanks in advance!

Benjamin

[markus kossmann]

One idea: Both CVI and your application use the same DLLs. If you start CVI first , these DLLs are  loaded into memory at that point but  if you start your application standalone they have still to be loaded from disk.

[Benjamin_C]

Hi Markus,

Thanks for your answer!

My own DLLs are loaded at the executable startup with static libraries. If you think of the DLL of the CVI Runtime, I supposed that they also are loaded with at the executable startup (with the InitCVIRTE() ?).

In addition, the decrease of execution speed does not only occur at the start of the application. My application launches operations in cycles. I always have the same execution duration, for the first cycle or the Nth cycle.

Thanks nevertheless! Of course, If you have another idea, don't hesitate to submit it.

Benjamin