initialisation on toggle debug mode on/off

Discussion in 'Building With Reaktor' started by MvKeinen, Apr 23, 2021.

  1. MvKeinen

    MvKeinen NI Product Owner

    Messages:
    181
    Hi!

    is it normal that initialisation happens twice when toggling the debug mode on or off, or is it a problem with my structure?

    thanks in advance!
     
  2. Paule

    Paule NI Product Owner

    Messages:
    7,397
    IMO it sends a GRE on or off.
     
  3. MvKeinen

    MvKeinen NI Product Owner

    Messages:
    181
    I now have the problem that the readout of an array has an initialisation problem.

    on init an array will be written via an iterator and after that it it is read by an other iterator. Both are serialized with an order module.

    When I start reaktor, and with a switch GRE all is as it should be. But when turning debug on or off the first 3 indices of the array have the same values as the indices 3,4 and 5. From 6 on the values are correct.

    It happens only on debug toggle.

    Wasn't there a pdf file with an in depth description of the different initialisations? Made by a user.
    Has r6 changed that?
    Where would you look into having this problem?

    thanks in advance
     
  4. colB

    colB NI Product Owner

    Messages:
    3,854
    Mark Wadewitz (Quietschboy) did some important research in that area - I think during the development of ACEW, and created a table in pdf format. There were some circumstances where core array data could be corrupted, or left in an unpredictable state. I'm not completely sure if there's a fully up to date version of this information though because initialisation did change in a relatively recent version update... maybe6.3? can't remember off hand.
     
    • Like Like x 2
  5. colB

    colB NI Product Owner

    Messages:
    3,854
    Can you upload an example ensemble that demonstrates this issue?
     
    • Like Like x 1
  6. MvKeinen

    MvKeinen NI Product Owner

    Messages:
    181
    I'll try to upload that tomorrow. It would take time. because its relatively deep embedded in the structure.

    thanks
     
  7. colB

    colB NI Product Owner

    Messages:
    3,854
    is it polyphonic or monophonic?
     
    • Like Like x 1
  8. Michael O'Hagan

    Michael O'Hagan NI Product Owner

    Messages:
    1,151
    These are all 4 of the documents I have regarding init events and their properties, If there is anything newer I'm not sure, some go back to R5 and one is from version 6.0.4, but this is everything that I know of, I hope this helps.
     

    Attached Files:

    • Like Like x 4
  9. MvKeinen

    MvKeinen NI Product Owner

    Messages:
    181
    its monophonic.

    Thanks a lot for the documents. Ill go through them now.

    I'm using the G out of the snap value array as a trigger for the iterations when it sends a zero. When I use both events from the G out it works. But at a later stage of the project that won't be possible. but i don't want to do the procedere twice.

    damn.
     
  10. Paule

    Paule NI Product Owner

    Messages:
    7,397
    One is in german language.
    One was the same Michael uploaded.
     

    Attached Files:

    Last edited: Apr 27, 2021
    • Like Like x 1
  11. MvKeinen

    MvKeinen NI Product Owner

    Messages:
    181
    user error :confused:o_O

    but there was definitely something undefined, because on further investigation i got it to a state where every 2 or 3th initialisation while in core it was good.... but jeah, my mistake was that between the iterations there was a lot of resorting of the array done and I resorted 2 things in one stage so that indices got messed up. now its all good. I passed the second half of the sorting from the read- stage to the write stage. no conflits anymore. I still dont understand why it sometimes worked on sometimes not. But learned a lot about initialisation. Now I just have to learn what "racks" is and I'm a reaktor master! :):)

    thanks to all! great forum btw.
     
  12. ANDREW221231

    ANDREW221231 NI Product Owner

    Messages:
    809
    it makes me a bit paranoid knowing how seriously people take initialization and how willfully i ignore it. i mean it's caused problems before so its not that i don't understand it, but most people seem to have learned a healthy fear of it
     
  13. Paule

    Paule NI Product Owner

    Messages:
    7,397
    Racks are presets, meaning that there are virtual cables between the blocks in their own area.
     
  14. colB

    colB NI Product Owner

    Messages:
    3,854
    :)

    It's not so much fear. Just that in certain circumstances, it can cause problems that are not initially (see what I did there?) obvious, but when you do eventually find them, not always easy to fix.

    Initialisation is complicated, and there are no official documents that explain all of what is going on. Undocumented behaviour is not a good thing in a programming environment, particularly when it is behaviour that is always there, and can break otherwise correct code.
     
  15. ANDREW221231

    ANDREW221231 NI Product Owner

    Messages:
    809
    i think you're right, respect is probably a better word. the fear is probably more applicable to me (still hyperbole) since its an unknown it could potentially be hiding anywhere, now respecting it makes a little more sense. probably some types of things get hit worse than others, maybe that's why i've been lucky.

    i solicited for anyone's worst initialization stories, but it occurs to me there's already a goldmine here already with a simple search. that will probably do

    suppose going back to far won't do. reaktor 6 came out in 2016? 2017?
     
  16. Quietschboy

    Quietschboy NI Product Owner

    Messages:
    544
    Hi Colin,
    do you have any information or source of what could have been changed for Reaktor`s Initialization since 6.0? I am not aware of such.

    Btw, an `unpredictable` state of Core Arrays happens when editing the Core Cell or when having Reaktor´s debug tool activated and then diving into/out of Core Cells or diving through it´s macro hierarchy.
     
  17. Quietschboy

    Quietschboy NI Product Owner

    Messages:
    544
    I am not sure if the Snap Module itself was your problem, but i just want to mention at t his point, that the Snap Module´s Initialization moment can differ from the main initialization moment. Means: Some Snap Modules Outports fire later. I don´t have the details in mind for now to be exactly, sorry. But all interested peoples can use ACEW to check that out.
     
    • Like Like x 1
    • Informative Informative x 1
  18. colB

    colB NI Product Owner

    Messages:
    3,854
    I probably misremembered that from the 6.0 release.