LabVIEW memory management changes in 2009-2011?

[Ben]

Yes.

The challenge when working with performance (memory CPU or other) is that to decalre ti abug we should show that the behaviour deiffers from the spec. In this case, there is no spec, so hard to decalre a bug.

We'll see what develops.

Ben

[Eric-M]

Hi Ben and Matthew,

Thanks for clarifying Matthew.  I've spent a good amount of time today looking into what may be causing the increased memory usage with the variants from 8.6 to 2009.  Ben has had the same reasoning as I, that this has something to do with the change in compiler.  I've been able to confirm the results of memory test 1 using LabVIEW 2011 on my computer, and I'm working to reproduce the results of the 8.6 results as well.  If those results are the same as in the table, then I will go ahead and file the CAR.  Either way, I will reply to the thread with my results.

Matthew, what OS are you using to run the VIs?

Thanks,

[Matthew_Kelton]

Eric,

I am using Win7 32 bit, but in 8.6 at least, the memory usage is basically the same on XP SP3.

[Ben]

---

@Eric-M wrote:

Hi Ben and Matthew,

Thanks for clarifying Matthew.  I've spent a good amount of time today looking into what may be causing the increased memory usage with the variants from 8.6 to 2009.  Ben has had the same reasoning as I, that this has something to do with the change in compiler.  I've been able to confirm the results of memory test 1 using LabVIEW 2011 on my computer, and I'm working to reproduce the results of the 8.6 results as well.  If those results are the same as in the table, then I will go ahead and file the CAR.  Either way, I will reply to the thread with my results.

Matthew, what OS are you using to run the VIs?

Thanks,

---

Eric,

Please share the CAR once created.

Matthew,

Please accept my thanks for giving us the "heads-up". Like when driving, I prefer to be a bystander. Thanks for taking the wheel a drivning off into the darkness.

Ben

[Eric-M]

Hi Ben and Matthew,

I was able to reproduce the results locally using both LabVIEW 8.6 and LabVIEW 2011, so I filed the CAR just now.  The CAR request number is 321353.

Thanks,

[Ben]

---

@Eric-M wrote:

Hi Ben and Matthew,

I was able to reproduce the results locally using both LabVIEW 8.6 and LabVIEW 2011, so I filed the CAR just now.  The CAR request number is 321353.

Thanks,

---

Thanks for the CAR!

Ben

[craige]

Does anyone have a good work around for this until the issue is resolved?

Thanks,

Craig

[Ben]

No I do not know of a work-around beside the obvious, Don't upgrade Don't use var...).

BUT...

With a CAR number you should be able to call NI and ask what is the work-around?

If there are work-arounds, it would be included in the CAR info.

Ben

[Matthew_Kelton]

Craig,

As I mentioned in my original post, dropping Request Memory Deallocation VIs helped this somewhat.  I couldn't get back to LV8.6 usage, but was able to get my code to release enough memeory to at least make it through the import without crashing.  The best place I had it was in the DB Toolkit VI, but of course, that will get lost at some point if I move the code to another PC or upgrade the DB Toolkit.

[Matthew_Kelton]

So, I have 2012 now, and figured I would update this thread with the info.  The issue has not been resolved.  All memory measurements done using task manager and are in kb.

	Open LV	Open VI	Run 1	Run 2	Run 3	Close VI
Mem Test	73848	73764	680456	1070388	1070388	1007132
Mem Test 2	73784	73800	590220	590220	590208	543376

	Open LV	Open VI	VI Run	Close VI
Mem Test 6	72288	193532	255544	242696
Mem Test 7	73868	132340	132340	126452