Teststand memory usage

[Alex_Haak]

When generating results teststand gradually uses more and more RAM. This happens when I generate a .tsr aswell as ATML(desired format). When i dont have any results being recorded the problem disappears, despite the same steps being executed. This indicates that it is the report generator that is causing the problem.

I have the following settings:

 - Model options

>> Discard results... [checked] (results are not being used anywhere else, other than default report generation)

 - Report options(for ATML)

>> Append If File... [checked]

>> On-the-fly reporting [checked]

>> Only display Latest results... [checked]

>> Processing interval: 9.50 Automatically tune interval [checked]

>> Maximum number of results: 30

Im using teststand 2012.

Please help!

Thanks

Alex

[Norbert_B]

Alex,

is this an observation or is it a problem (so running out of memory after some time)?

Do you observe this behavior also for other report formats like HTML/ASCII?

---

@Alex Haak wrote:

[..]When i dont have any results being recorded the problem disappears, [..]

---

Do you disable result recording or simply disable report generation?

Is the behavior observable with any sequence or is it connected to a specific one?

Norbert
 

[crossrulz]

This sounds like you are disabling results when you don't have reports.  Basically, the Results array is increasing in size for each step that has a result is ran.

[Alex_Haak]

This is a problem. Currently the RAM is running out after approx 14hrs and it needs to be able to last 72hrs.

The same happens with HTML, at a slower rate.

I disabled report generation rather than just disabling results.

Thanks

[Anand_Jain]

Hi

To help narrow down this issue can you please post an example sequence file which reproduces this issue and all files under C:\ProgramData\National Instruments\TestStand 2012\Cfg\ModelPlugins

Thank you 

Regards

Anand Jain

National Instruments

[Alex_Haak]

I dont have access to the computer right now, but i havent got any plugins and the sequence file basically has a number of dll calls and a call to another sequence which also has a number of dll calls. So im fairly sure the dll isnt leaking any memory as without report generation the problem disappears.

I seemed to have fixed the problem yesterday afternoon, but...  i seemed to have changed something because it is back!

Could it be something to do with the way i am loading/unloading the other sequence and the dll that my sequence uses?

otherwise maybe the advanced result processing options?

another problem could be that i am filtering my results... can i be sure that teststand discards the results that i am not recording?

I think these are the only things i changed.

Thanks so much,

[Alex_Haak]

Another thing could be that im not including anything under step results. Instead i am using additional results.

[jhedr]

I am seeing the same issue, it appears to be something with the Number of Results but so far I have not pinpointed it to that.  Also using TestStand 2012 SP1, and started seeing the problem when we added a loop to check BITS (could add upto 430 results) depending on the Bit stream.  We are performing Thermal Testing, so it does multiple loops through the code.

I turned on Discard Results and Display only Current with On-The-Fly Reporting, but so far I am still seeing the memory usage increase.