Dirk J.

QUOTE(alnaimi @ May 8 2007, 06:11 PM) no need to shout. the code you uploaded does not clarify what you want. what do you want with the for as far as I can tell, you just duplicated your code into a second frame. if I understand your question, what you want is shown in the ...\National Instruments\LabVIEW 8.2\examples\general\graphs\gengraph.llb\XY Graph.vi example.

QUOTE(alnaimi @ May 8 2007, 04:18 PM) sigh... again your code is broken. as I told you before, most of us do not have the KIDAQ subVI's.

hey grasshopper (you're going to regret choosing that name.... ) in principle re-sampling of your data, horizontally and vertically, shouldn't be that much of a problem. can you share some of your code, so we can see to what point you got? -sensei

QUOTE(hooovahh @ May 7 2007, 04:08 PM) Actually, I think you are flaming. Labview code is more readible if it flows from left to right, but as ActiveX and .Net nodes tend to take up a lot of screen real estate, it is sometimes more convienient to have nodes placed topleft to rightbottom. So sometimes, data will flow from right to left. An alternative would be to hide full names, but that would decrease the quite extensive self-documentation of these nodes. And there's always MSDN. You can always 'clean up wires' and use highlighted execution if you're in doubt. If it doesn't open up in your LV version, was made in a newer one. Cosmin provided a quick solution to the problem, not submit something to the code repository.

QUOTE(alnaimi @ May 4 2007, 01:46 PM) alnaimi, stop shouting. 1) I do not have your KI AI Sample Channel available. 2) I replaced it with some code to generate one value each time that VI was called. See the attached file.

QUOTE(alnaimi @ May 4 2007, 01:20 PM) they are there on your system, not on mine. simulating is just replacing the missing KI AI Sample Channel with some number-generating VI in order to build the graphs.

QUOTE(alnaimi @ May 4 2007, 12:39 PM) Maybe because your VI is broken (the KI AO xxxxx, and KI AI xxxx VI's are missing) But, if I replace the KI AI Sample Channel VI with some simulating VI, I just get 3 graphs, so I can't reproduce your image....

QUOTE(Michael_Aivaliotis @ May 4 2007, 12:13 PM) Since you know which posts this concerns, do you have the feeling that the mistreatment had to do with the topic of their question or problem, or with the way it is asked (complying to forum etiquette)? QUOTE What I am concerned about is that as a result of this discussion, we now have long time members doubting if they are advanced enough. What is the criteria for advanced? Can someone define this? For example: if you are able to communicate about your problem, as in "I'm getting error code 15 when running this DLL", perhaps with some more detail or code snips, then you're advanced. If you "need solve projec str8away, pleazzzzzzzzzzzzzzzzz u help", then you're not. Regardless of the topic.

It's not so much about newbies, but about homework hustlers. In my opinion the most annoying HH posts share two common characteristics: 1) no evidence showing understanding of the actual problem (that needs to be solved using LV) 2) SMS/Chat-type spelling. Both points don't really relate to knowledge of Labview nor to English language skills. Even with limited knowledge of a programming language an (engineering) student should still be able to analyse the problem and come up with some sort solution or algorithm. If they do, most if not all members will be happy to constructively comment. Regarding the second point: it's hard to take such posts seriously. It has nothing to do with English language skills. These posters are able to read this forum (and we all asume they searched and studied it, before posting their question) so they have a huge amount of examples showing them how to effectively ask a question. So being a Labview newbie using "charcoal english" (is that term only used in the Netherlands?) is not a problem, as long as you communicate about your problem -and more importantly about your approach to solving- it in a manner that is conventional to the forum.

QUOTE(tcplomp @ May 1 2007, 07:01 PM) The typecast function doesn't accept 2D arrays as input (see image). I grab the data as 1D array and at some point need to convert it to a 2D array (image, actually), so its a reshape from 1D to 2D.

QUOTE(tcplomp @ May 1 2007, 07:01 PM) on my laptop its about 1.4 times faster. I didn't know about the typecast; I'll try that, thanks.

Better late then never... I'm attaching a project showing my LV implementation of the Fastest Fourier Transform of the West. The example implements a complex-to-complex transform as outlined in the fftw manual (which is also included). The example VI is documented; the text is reproduced here. In FFTW, you create a "FFTW plan" (a pointer to a structure containing all relevant information) by calling an initialization routine. In the present implementation, the planner takes the FFT size, and a planning method as inputs, see the Block Diagram for details. The plan is optimized for /your/ hardware configuration. Depending on the method you choose, this can take quite some time so you typically use this 'outside the loop'. You can then execute the plan as many times as you like on different arrays by calling the execution routine inside the loop. Afterwards, clean up by calling the destroy routine. This example shows a 1-dimensional transform. 2D, 3D or actually any-D transforms are also possible by calling different initialization routines; see the FFTW documentation for more details. It is also possible to perform FFT's along for example only 1 dimension in a 2D array, which was my origional problem. You accomplish this by calling yet another initialization routine (for this example, you would configure the CLF node to call fftw_plan_many_dft with the "howmany" parameter set to 10. The input array would then be a size 20480 1D array, and you wouldn't need a FOR-loop). I'm still working on this one. The way I have it working needs to reshape the arrays each time which proved to be slower than calling the 2048 array FFT 10 times. The planner routine has a direction parameter; a value of -1 gives the forward FFT; a value of 1 the iFFT Of course you are not limited to to complex-to-complex transforms; real-to-complex; real-to-real, ... etc is also possible, again by calling a different planner library function. See the FFTW manual for the details.

Lab, before asking your next question please read the forum guidelines first. You will learn that in general, forum users are not inclined to solve your homework projects for you. At the very least, try to provide some example code. Nullll, if you have the solution to the problem, by all means post it. We're here to learn from each other

or this underestimated tool...

Yes, I got it to work. I'll post example code sometime this week. It may take a couple of days though, I'm quite busy at the moment (implementing the FFTW in my app :-) and it really needs some documentation because it is not 100% straightforward from the wiring what's going on. In the mean time, if you're really interested, I'd suggest you read the whole FFTW documentation from the site, not just chapter 2, because that will clarify a lot of what's going on in the code.

This worked for me: move the tms_reset DLL call into a sub-vi, see screenshot. The error message still doen't make sense to me though... good luck. (found out by systematically removing stuff from the BD)

You could alway go through the windows API calls, but that's usually a bit troublesome at best. An easy workaround is to create an internet shortcut instead, which is just an *.ini file, with *.url extention. The plain format is as follows: [internetShortcut] URL=c:\windows IconFile=C:\WINDOWS\system32\SHELL32.dll IconIndex=4 HotKey=0 IDList= See attached VI for details...

QUOTE(Tomi Maila @ Mar 6 2007, 08:54 PM) that triggers me... i'm using firefox w/greasemonkey, writing a script to 'downscale' all text shouldn't be that difficult.

Hi Sherif. I hope you will find the forums useful, and that you'll make contributions that will benefit other people as well. Please don't be offended by my next remark: in my opinion it will help if you don't use too much font-formatting or emoticons; this will increase the "readability" of your posts. That will increase the willingness of people here to help you. Best regards, Dirk.

QUOTE(dsaunders @ Mar 1 2007, 05:22 PM) It is a big annoyance.

I have a similar question... how to handle functions from a dll that pass a pointer? I'm trying to use the FFTW routines, a code mockup looks like this: 'plan' is of type pointer plan = fftw_plan_dft(datalenght, inputarray outputarray, direction, method) .. fftw_execute(plan) .. fftw_destroy(plan) how is 'plan' represented in LV, and how do I configure the library node in this case?

Sounds impressive (and difficult). At the moment I don't have the right processors...

I assume it works in a "first call" kind of way, to initialize the 'tbl' input of the FFT library node. There are 2 mysterious inputs to the call: this 'tbl' and an 'org' parameter, both integers. Preallocating an array increases the speed of the above example with about 20% !!

Try disabling debug mode on your VI's, saving all them and running to see what sort of speed increase you obtain. -- did that. Make sure you are not using any property nodes -- none there. You could also play around with multi-threading, priorities, re-entrancy, and code parallelism to speed things up further, but that is a topic for another day! -- have been doing that, but not to full extent I guess. PS. Posting code is a good way to generate a higher quality discussion. -- I know, but so far testing has been pretty simple, see image of bd. At this point, I don't have any experience with using external FFT code, so nothing to share yet...

1 Which version of LV are you using? I think Version 7.1 or so there was a major re-write of the math functions. --- 8.20 (faster than before, I agree) 2 Make sure you don't have any coercion dots (data type conversion) into and out of your FFT routines. This will slow things down. --- Has been taken care of. 3 For best speed even with newer hardware, desktop version of the processor is faster than laptop. I have a Dell Inspiron 9400 laptop, and a Dell Dimension 9200 desktop; the desktop is much faster than the laptop. --- Laptop is not a strict requirement. 4 A built application (exe) will give you slightly better speed than the code. --- Can you quantify that (approximate percentage or so)? AppBuilder has been ruled 'too expensive' up until now... -thanks