erratic slowdowns on Win 7 machines

[dacad]

Due to corporate requirements, all the company pc's must be updated or replaced with Win 7 systems.  😞

The first system I completed began life using LV 2011 with Win 7 and for the most part, things work pretty good though there are some glitches.  Once in particular will show up somewhat erratically with no pattern. Sometimes can go days or even weeks and sometimes twice a day.

It manifests itself by what appears to be a sudden pause in the execution of the system. Certain functions will recover but there is always something behind the scenes that will quickly cause things to not work.  The only visual indicators that I have observed is that the system time that I display on the UI stops updating. It usually ends up requiring an application shutdown and restart and occasionally a system reboot.

I had always thought there was something in my code that was driving this but after this week, am re-thinking that.  I have another application that has been running quite a long time on a Win xp box in anything from LV 8.6 to 2011.  Never a problem.  I have ported the app over to a brand new dell with win7 and loaded with LV 2013 (w/ update).  All of the sudden this thing will exhibit these strange pauses quite often.

It is much more noticeable on this system as it performs screen updates after each of the 12 or so individual tests that run consecutively. Once it wigs out, it never recovers even though it will continue to run through the tests, albiet very slowly.  The screen updates lag and cause the entire sequence to go from a three and a half minute sequence to a 12 minute sequence, give or take.

Before I go posting screen caps and code and all, I just wanted to see if there are any known issues when updating to a win 7 box (is 32 bit).  It seems to be centered around screen updates and perhaps global variables maybe ?    I keep my gv's to a minimum.   I run a basic four loop parallel architecture (Event Structure loop, screen update/global var loop, user event/ui state machine and, the execution state machine (does all the grunt work).

Are there any code issues known to not get along with win 7 ?     Win 7 display issues ?    global variables ?   constant time polling (45 ms delay in loop) ?   other ?

Thanks for any thoughts on this.

Doug

[JÞB]

Turn off Aero themes!  Those themes do some nasty graphic magic that can send the OS tempoarrilly out to lunch and has been known to inhibit expected panel updates.

[dacad]

One of the first things I do before even loading the first vi is to switch to the windows classic theme.

[JÞB]

---

@dacad wrote:

One of the first things I do before even loading the first vi is to switch to the windows classic theme.

---

OK so not an AERO theme issue. 

Next,   You did confuse me a bit on what you are putting on the new win7 machine. How are you distributing your "App"?  Did you mass compile an existing project and rebuild it in 2013 or are you running source code?  Are there dynamic calls in the dependancies?  Are there calls to files that exist in folders that changed betewwn XP and 7? (GAC changed somewhat)  Are there dependancies in NI\Resources? (LabVIEW 2012 replaced the absolute with a symbolic path) were any search paths options carried over from LabVIEW x.x to LabVIEW 2013' LabVIEW.ini file?

[dacad]

Had run in LV environment and as exe on previous machines.  Starting out by running in LV programming environment to get any bugs worked out on new machine.

No dynamic calls.  Outside file calls are to fixed paths controlled by ini file. 

Have been running app solid for past hour, no stopping, just running same unit over and over.  Glitch occured twice in 18 test sequences.  Without adding something to the code, I don't currently have a method to determine exactly where the glitch occurs.  Problem of course is that I also don't know if I can trust the results despite what the returned values end up being. And, when the test takes 10 or 12 minutes versus 3-1/2, productivity takes a big hit.

Nothing in my test sequence changes. It completes the same operations each time. Being that the glitch occurs erratically, I can't but help believe it is related to some bad interaction to the OS.

I guess I could try stripping functions out of a copy and see when it quits happening.  Is there some type of tracking code or analysis tools I could use to monitor the app while it is running.  I have never used any of the vi analysis tools so am not real familiar with them.

Thanks.....

[dacad]

One thing just as an fyi.  I am using a legacy pci-6014 daq.  Not sure if that would have any impact.  MAX picks it up ok.  Other than that only hardware is a xantrex/ametek/sorensen power supply connected with a usb cable.

[JÞB]

Those Sorensons can be a B!^@h.

First, make sure the driver is mass compiled.

Are you opening and closing multiple IO Sessions?

Do you have the windows power options configured for the HUB?

Is the VCP driver up to date?

Is the USB HUB self powered from a reliable wall wart? (Dell USB Port power supplies are notorious in my experience)

Is the enviornment "Noisy"? and if so (assume it is) is the USB Cable built with quality shielding and ferrites? 

[dacad]

Yea, sorensens can be troublesome sometimes but I found that once you get them talking, there ususally are not any other troubles to deal with.

pwr sply is connected straight to computer usb. No special setups or anything.  I talk to the power sply a few times during the test sequence and get responses and would clearly expect a com error if there were issues there.  Only opening one instance with it and closing it on exit.

Have been playing with the code that impacts screen updates and have been able to reduce if not eliminate the glitch.  Certainly at least affecting it. Disabling certain display functions makes it go away (at least for the limited duration I have been able to test within) or at least reduces the impact.

I am convinced there is some display issue going on.  Bad part is, I may be able to band-aid it but not knowing the root cause means it could haunt me forever.

Who knows, could be a conflict with some corporate installed watchdog utility.   The fact that I found little to nothing related to this on the forums means it must be somewhat specific to what I have going on. That or everyone else who experienced it found the solution much faster than I.

If I do come up with anything definitive, I will certainly append this thread accordingly.

Onward I go.....

[0utlaw]

Hello Dacad,

Just to check- are you using DAQmx or the legacy NI-DAQ with your ?  I don't believe that there was ever a full release of NI-DAQ for Windows 7, as per this KnowledgeBase:

Traditional NI-DAQ (Legacy) for Windows Vista, Windows 7, and Windows 8

http://digital.ni.com/public.nsf/allkb/8EC9E81C3DE4E619862573930083BD7A

This may or may not be affecting your application as it sounds like you've already narrowed things down to a UI bottleneck of some sort, but is worth keeping in mind.

Regards, 

[Yamaeda]

Assuming it is a graphic issue, do you change visibility and transparancy on objects? 

Is it random time when glitching out, or only after having run some time? (Could it be some race condition or resource that gets eaten up)

/Y