ShaunR

2 hours ago, Bruniii said:

Thanks for trying!

How "easy" is to use GPUs in LabVIEW for this type of operations? I remeber reading that I'm supposed to write the code in C++, where the CUDA api is used, compile the dll and than use the labview toolkit to call the dll. Unfortunally, I have zero knowlodge in basically all these step.

There is a GPU Toolkit if you want to try it. No need to write wrapper DLL's. It's in VIPM so you can just install it and try. Don't bother with the download button on the website-it's just a launch link for VIPM and you'd have to log in.

One afterthought. When benchmarking you must never leave outputs unwired (like the 2d arrays in your benchmark). LabVIEW will know that the data isn't used anywhere and optimise to give different results than when in production. So you should at least do something like this:

On my machine your original executed in ~10ms. With the above it was ~30ms.

Nope. I can't beat it. To get better performance i expect you would probably have to use different hardware (FPGA or GPU).

Self auto-incrementing arrays in LabVIEW are extremely efficient and I've come across the situation previously where decimate is usually about 4 times slower. Your particular requirement requires deleting a subsection  at the beginning and end of each acquisition so most optimisations aren't available.

Just be aware that you have a fixed number of channels and hope the HW guys don't add more or make a cheaper version with only 2.

Post the VI's rather than snippets (snippets don't work on Lavag.org) along with example data. It's also helpful if you have standard benchmarks that we can plug our implementation into (sequence structure with frames and getmillisecs) so we can compare and contrast.

e.g

3 hours ago, Rolf Kalbermatter said:

I have so far not found a way that makes those paths automatically fixup at package creation, since the path seems to need to be absolute

This is why it takes me hours to make an ECL build that works and is one of the many reasons only Windows is now supported (can load from same dir). Even then. I have to fight VIPM to get things in the right places.

I refuse to do #2.

Well. There's a few problems but the reason it's not showing the next image is because you increment the counter during acquisition until it's 300 and when you start the next acquisition it indexes into the path array at 300 (which yields Not A Path).

You've confused your 30 second timer with the file index. Make a proper timer with a time function and increment the index on stop.

17 hours ago, Jordan Kuehn said:

Not to derail the above conversation, but can this ping function reliably time out in Windows sub 1s?

Yes but you'd need to define "reliably". I think there would be  few milliseconds of jitter and probably a minimum of about 4ms.

There are surprising few situations where a parallel for loop (pLoop) is the solution. There are so many caveats and foot-shooting opportunities even if you ignore the caveats imposed by the IDE dialogue.

For example. For the pLoop to operate as you would imagine, Vi's that are called must be reentrant (and preferably preallocated clones). If a called VI is not reentrant then the loop will wait until it finishes before calling it in another parallel loop (that's just how dataflow works). If a called VI is set to reentrant shared clones then you get the same problems as with any shared clone that has memory but multiplied by the number of loop iterations.

Another that you often come across with shared, connectionless resources (say, raw sockets) is that you cannot guarantee the order that the underlying resource is accessed in. If it is, say, a byte stream then you would have to add extra information in order to reconstruct the stream which may or may not be possible. I have actual experience of this and it is why the ECL Ping functionality cannot be called in a pLoop.

I'm a simple programmer.

Catch them all where possible (try { /* */ } catch (...)) and pass errors back as a return value.

That's my tuppence.

23 hours ago, jhoskins said:

Nice VI by the way, mind sharing

It's the example from MDI Toolkit for LabVIEW

16 hours ago, Jim Kring said:

Here's a new take

Here's slightly different take on that idea

>> LabVIEW Idea: Enumerated Variant <<

It works with, but is not limited to, classes.  It's inspired by languages like Rust (enums) and Zig (tagged unions)

If you put a Call By Reference Node in the case you've reinvented Dynamic Dispatch with extra steps.

1 hour ago, Rolf Kalbermatter said:

With a little extra effort in the Base Class implementation for a proper factory pattern implementation,

Booh with bells on.  Hello 7 hr build time. 

41 minutes ago, Rolf Kalbermatter said:

I may revisit my earlier  attempts of writing a fully LabVIEW OOP based solution for this

Booooh! 

1 hour ago, Rolf Kalbermatter said:

You really may be stressing LabVIEW's window management capabilities beyond reasonable borders with so many subpanels present at the same time

Agreed. Even just 100 and UI updates become a bit flakey.

8 hours ago, fabric said:

I've enjoyed this hack for many years, but noticed it is not working in LV2023Q1. 

See here for problem description: https://forums.ni.com/t5/LabVIEW/Darren-s-Weekly-Nugget-05-10-2010/m-p/4360614/highlight/true#M1280554

They have fixed a bug, is my guess. Concatenating by using null char is a huge security smell.

Multiple file types are [supposed to be] defined by using the semicolon separator. Does "*csv;*txt" not work?

10 hours ago, JKSH said:

What communication protocol(s) does your digital X-ray plate use to receive commands and transmit images?

It's probably DICOM which is partially supported in the Biomedical Toolkit.

On 3/9/2024 at 10:36 PM, Mahbod Morshedi said:

I was just wondering about the "array with tag" or "Var attribute" since there is not much information about the latter. I am unsure if the data is more susceptible to corruption, gets lost, or if there are other complications I do not know about. 

The variant attribute cannot have duplicates. If you set it, it overwrites the previous values. With the "array with tag" you can have duplicates.

Try

print(json.dumps(convert['test'], ensure_ascii=False))

1 hour ago, David Boyd said:

I hope you corrected the reentrancy settings for the two malleables in the library when you backsaved to 2021.

Nope. I just provided the back-save of your submission.

20 hours ago, X___ said:

Can this been downgraded to 2021 or earlier?

Here you go.

CRC-LV2021.zip

7 hours ago, David Boyd said:

are you thinking that "shared" (vs. "preallocated")

Yes. that's exactly what I am thinking (but poorly communicated). This is a common known gotcha for VI's with shift register memories (not the first call primitive per say).

It will probably only bite you when you have multiple instances and where it's being used with different CRC types with different integer lengths. 

Here's an example:

sub VI set to Preallocated (what we expect-11 more than the intialise value)

sub VI set to Shared:

If you run continuously, you will see other values as different threads become available at different times.

rentrant clones.zip

Nice.

This is probably one of the only times (1 in 1,000,000) I would suggest an xnode may be preferable-and specifically for the lookup tables which should always be more performant.

With an xnode, one can pre-calculate the tables at design time based on the type and save the cost of generating the table at first run run. This will also mean that the calculation will be constant time whether first called or later. xnodes are tricky and complicated beasts so I could understand not wanting to go down this hairy rabbit hole littered with rusty nails.

Speaking of the table generation; I noticed you have the VI set to reentrant clones. I think this will be  a problem when you run multiple instances as the shift registers may not contain the values you expect per instance.

19 hours ago, hooovahh said:

Oh but I did look up the 2023 roadmap, and unicode support was changed to "Future Development".

My takeaway from the roadmap is that they are concentrating on interoperability and relegating products to backend services. This was the direction NI were taking before Emerson but  expect it's taken on a new impetus since the takeover. I expect the awful gRPC to be leading the charge so they can plug the NI products to their products. Unicode support isn't a consideration for that since the UI will be elsewhere.

The ini (or config file as NI calls them) are only written when the reference is closed.

You could put the flush in there but I don't think that's your problem.

I have seen this before and it is usually on a failing hard drive or one that is about to fail. NTFS has a sort of self healing which is insidious for small files (less than the cluster size) and this is the usual result of it "healing". Larger corruptions are more detectable with chkdsk.

Quote

NTFS continuously monitors and corrects transient corruption issues in the background without taking the volume offline. This feature is known as self-healing NTFS, which was introduced in Windows Server 2008.

(source)

I would take a look at the S.M.A.R.T (Self-Monitoring, Analysis, and Reporting Technology) information of your disk and pay particular attention to the write failures/errors.

17 hours ago, dcoons said:

also a coworker pointed out those scroll bars are likely in the same boat 😭😭

..and splitters.

It's these sorts of issues that are the reason some of us abandoned the LabVIEW UI altogether.

There is an "Element" property for arrays. You can get that then cast it to the element type. Then your other methods should work (recursion).

Boolean Get Image Testing_sr.vi