Large string array usage speed

[Nallez]

I'm using a large (250x17) string array and reading certain items from it to a cluster.

The items taken is maximum 3x20 at one time to the cluster which shows them.

I just have one problem:

To show these 3x20 values in cluster, it takes considerable amount of time (1.8seconds!)

Is there any better way to read from String arrays than just index them?

[altenbach]

Is the string array 2D with 250x17 elements? How long are the strings in each array element?

You mention also a cluster, but don't provide any details.

Can you show your code?

[Nallez]

[altenbach]

I don't see the "cluster", just tons of value and other property nodes.

Some real code would be much more useful than a picture, especially since large parts of the code are outside the image boundary. Where do the arrays of references come from? Is this a subVI

[Nallez]

Array is 2D and reference comes from file.

Cluster has 1+18*3 elements and it comes as a reference from the main VI.
Cluster comes from the upleft (the green wire :))

Only array loading outside image boundary.

I'd have to release the whole code to make it work properly and that is not possible.

But if there isn't anything visible wrong with the code, then I just have come up with something.

Message Edited by Nallez on 04-27-2007 01:12 AM

[parthabe]

We re handling a 2D Str array of Nx40plus size in our project, no problem till date, when converting a to cluster.

Maybe, as Altenbach has mentioned, you need to post the small part of the exact conversion code part (without references, replaced with controls & indicators) to shed more light on the unncessary delay caused by the conversion.

[GerdW]

Hi Nallez,

one possible reason:
when using property nodes LabView tries to update the corresponding frontpanel. This may take a while when you do this for a whole array of values...
Try the panel property 'Defer panel updates' before and after your massive property node accessing
Or even better: try to use (much) less property nodes...

[Nallez]

Thank you GerdW!

The cluster properynode was the slowing part!

After I toned the property nodes down to 30% of the start, the VI runningtime became under 100ms

Message Edited by Nallez on 04-27-2007 07:33 AM

[GerdW]

Hi Nallez,

one more comment:
you can resize the 'Index array' function to give you more than one output. (See the context help.) This way you can get rid of a large amount of those functions too!

[Nallez]

Yes I tried that too beforehand, but it didnt give any positive effect on the performance.