Labview 8.5 Slow

[Kasper]

Hello

Anyone have an ide why Labview 8.5 is so slow ?

Kasper

Message Edited by Support on 10-17-2007 02:24 PM

[altenbach]

Obviously, we can only help if you are a bit more specific.

LabVIEW is very fast for most of us, so something is probably different in your situation.

• What is slow: opening a VI, editing a VI, running a VI, etc.
• What does your program do?
• A VI that just got converted from a previous version is typically slower and uses more memory until it it saved in 8.5. Keep that in mind when opening an old VI.
• What are your PC specs, CPU, memory, etc.

[Kasper]

> - What is slow: opening a VI, editing a VI, running a VI, etc.
Editing VI, open probertied boxes,saving etc. It's an old 8.0 Vi converted
to 8.5



> - What does your program do?
Log data on serial line and show on screen


> - A VI that just got converted from a previous version is typically slower
> and uses more memory until it it saved in 8.5. Keep that in mind when
> opening an old VI.
>
> - What are your PC specs, CPU, memory, etc.

AMD 4200 X2, 2GB RAM, 2x10K HDD in Raid

Kasper

[altenbach]

Different words mean different things to different people.

What is "slow"? milliseconds, seconds, minutes, hours, never?  ... or just slower than in 8.0?

Can you tell us a little bit more about your installation. Did you just upgrade to 8.5 from 8.0? Did you install LabVIEW in its default location or possibly on top of the previous installation? Is LabVIEW installed locally or in a network location? Does the previous version of LabVIEW still work?

Were there any errors during the installation? Have you tried to repair the 8.5 installation?

What else did you install (drivers, toolkits, etc.).

[crimolvic]

Hi altenbach, I have expereienced the same problem and I'll try to be more specific.

I have yesterday installed the version 8.5, before I had the 8.2. In this previous version I have done a vi which had principally a for loop where I perform several interpolations on a signal.
Example: on a 8 second signal sampled at 51200 S/sec the number of interpolations (or N of the for loop) is 411684. This process took me up to yesterday about a second.
To give you an idea, I run the vi with the same signal today with LV 8,5 before I start to write this post, now it rounds the 15000 iteration!!!

I did not upgrade the version. I unistalled the previous completely and installed the new one. LV is installed locally. There were no errors during the installation. I installed this time several toolkits and the NI DAQmx 8.50f3

PS: now it rounds the 30000 iteration ;(

Any idea which the problem could be?

thank you

[altenbach]

I don't quite understand what you mean by "rounding".

Do you mean it now only executes 15000 or 30000 iterations instead of 411684? Can you attach a simplified version of the VI and some data?

What form is the data (array, waveform, ...)?

Do you load it from a file (since you say you use the same data)?

How do you interpolate?

If I understand you right, you don't have problems with speed (as the title of this thread might suggest;)), just with the iteration count, which is different in 8.5

[crimolvic]

Hi altenbach,

thanks for replying. By "rounding" I meant that I excecuted the vi before start writing the previous post, and after let's say 1 minute the i counter of the for loop had a value of (just!) 15000; after 2 min, a value of 30000. It was just an example of how slow the vi is running.

Yes, I am reading the data (waveforms) from a file. I am attaching a simplified version of the vi with some data on it (is not the same data but the problem is the same). With this data, in one minute the i-counter has a value of 35000 (of 177131 iterations!)

As you can see I do have problems with speed, as the for loop is running too slow in comparision with the same vi on LV 8.2.

I did the upgrade because I needed to study some vi's from a toolkit, but if I can't solve this speed problem I will need to go back to LV8.2. The problem is that the vi's were already opened with LV 8.5, so I'll have to save all the (opened) vi's for previous versions...

Could you explain a little bit more the interation count differences between LV 8.5 and 8.2?

Thanking in advance,
CJMV

[Matthew_Kelton]

I saved your posted VI back to 8.2, and it ran in about the same time as it did in 8.5.  Are you sure you have made any changes other than the upgrade?

You should always make a backup of your code before upgrading to a new version of LabVIEW.

On my machine, the code you ran did not run as slow as the numbers you were giving, but did not run in one second.  The code you posted was broken, as it was looking for a missing subVI.  I replaced it with the other VI in your ZIP, which was not a direct drop-in replacement, but looks to do what you want it to, except it doesn't start at the last index (which would make it run slower).

You have some logic errors in your interpolation I believe, as some of your index values appear to be off.  Regardless, I posted a new rev of your main VI which runs significantly faster on my machine and uses standard NI array functions to do the interpolation.  Altenbach will probably have some additional or better suggestions to speed it up more.  I still don't think it should be taking as long as it is now.

[crimolvic]

Hi Matthew,

I don't know why you found the code was broken. There are only two vi's being one called by the other inside the loop.

However, you are right. your code does the same and in took less time to do it. With the data inside the vi (that is 177131 interpolations, or runs of the for loop) took my 36 seconds from start to stop. How much time does it take in your computer??

I think that's still to slow and it is probably due to something else, but I don't know what. Previous the installation of LV8.5 I completely uninstalled all NI software. I am ataching a screenshot of what MAX says I have installed at the moment.

Can it be something related to the Run-time engine?

[Matthew_Kelton]

It took about 8 secs on my computer.

I don't think the runtime engine is your issue.  If my PC is faster than yours, then I will run faster than you.

The big issue is that you are doing a large number of interpolations on the data.  Since your input data is not evenly spaced, you cannot quickly resample it.  Since your data is not "pretty," you cannot do simple fitting to determine your new data.  I am hoping Altenbach will have a different approach which will speed things up.