Full runtime engine versus Instrument support only

[GAINZT]

Hi all !

Since more than 2 years we've used more than 30 third party VXIpnp driver under our TestStand / LabWindows/CVI test systems.
From our own experience, we found that all drivers using the "Instrument support only" engine DLL (INSTRSUP.dll) where more subject to create memory corruptions after several sequence executions.
Typically, rebuilding drivers DLLs with "full runtime engine" instead of "instrument support only" always made the memory corruptions disappeared.
Removing the related drivers from our DLL module called by the TestStand sequences also made those corruptions disappears.
However, we were never able to explicitly demonstrate this behaviour in a "simplified" environment (i.e. without passing th
rough a TestStand Sequence etc.).

Since most recent VXIpnp drivers are compiled with "Full runtime engine", I supposed other people may have seen problems like this before ?!?

Please tell me your own experience on this !

[bilalD]

hello

I havnt really seen anything like this before. What did you mean by memory corruption? We havnt really seen anything like this happen before, and there is nothing much we can do if we cannot reproduce the problem. If you have some more information on where the problem is occuring, that would help too.

Thanks

Bilal Durrani
NI

[GAINZT]

Hi,

We actually have difficulty to isolate the issue.
What's sure is that after several executions, TestStand is unable to load DLLs with symptoms like :
1> Unable to load a funtion from our application DLL. Even if this function still exist.
2> Error in PostSubStep of the LimitLoader step.

Under TestStand 1.0.2 symptoms occured after less than 10 executions while using more than 5 drivers.

I can send you the full configuration of our development station by mail if you want but I'm affraid I can't post it here.

Sincerely,
Rémi THEVENIN