Jump to content
Jagware

Orion_

Level1
  • Content count

    1,073
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by Orion_


  1. ok, it has been about 2 months released in private beta now I try a public release :)

     

    here is a little Jaguar GPU Debugger I made to get rid of headaches I had coding on the jaguar gpu ^^

    it's far from being complete but it helps debugging the common gpu/dsp code I do, and hopefully will helps you too ;)

    the source code are fully included and can be compiled with the free "delphi like" compiler: Lazarus

    so you can add any features you want or correct some bugs ;)

     

    Enjoy :)

     

    GPU Debugger v0.8 beta, source included

     

    gpudbugv05.jpg


  2. Hi there !

    I just modified a bit the Virtual Jaguar emulator source code to support the loading of BIN file, in BJL format (i.e: loading at 0x4000)

    Virtual Jaguar is a slow emulator but a lot more accurate in emulation than Project Tempest !

    I hope it will be useful to developers :)

    I saw that the emulator also has a GPU execution log function (can be used like a GPU debugger !!!) you just need to change the doGPUDis flag to true in the gpu.cpp file and recompile the emulator !

    it crash most of the time after that, but we don't care, the only thing interesting here are the debugging info ! :)

    I always dreamed of a GPU debugger or emulator, know I have one !

     

    vj_bjl.zip


  3. I agree :)

    Emulators are great to test your code very fast, but it's also a trap that you can fall in, you start with a basic code that work well on jaguar and emulator, then code and code for hours using only emulator, it works perfectly, then try your code on a real jaguar and it doesn't work anymore. the fun thing is to find where is the bug because on the emulator it works just great !

    this happened to me some times ;)


  4. How many cycles per DSP instruction?

    one (but it's like on the gpu/dsp of the jaguar, you know pipeline ..)

     

    what will the price for the final jagcf be ?

    we don't know, SCPCD didn't ask for an Estimate at the company who will make the cf, because the final version of the cf is still not finished.

     

    Will there be list to signin to be able register for a first developers version?

    there will be limited registering with already reserved places ;)


  5. comme y'a toujours pas d'update sur le site du jagcode, tursi a poster son jeu:

     

    le site de tursi: http://harmlesslion.com/games/martian

    le lien jsii: http://jaysmith2000.ipbhost.com/index.php?...mp;#entry191355

    la video du jeu de tursi:

     

    c'est moi ou jsii ne peu pas s'empecher de faire soit des graphismes soit des bruitage de jeu 2600 ? :D

    bon et sinon son jeu .. personnelement, le fond est pas mal, mais le reste :/

    (et alors les bruitages >_<)

     

    a noter qu'il a réussi l'exploit d'avoir un fichier plus gros que le miens alors que j'ai 2 image plein ecran + 3 extrait de musique + bruitages :D (et sans compression)

     

    Zerosquare > si je t'envoi le bin version final de mon jeu, ça serais possible d'en faire une capture video ?

    histoire que je puisse poster sur youtube aussi :)

    parceque j'ai bien une carte d'aquisition mais j'ai qu'une sortie peritel sur ma jag :(


  6. bon ben on est que 2 :D

    y'a tursi et moi :/

    je vous livre une exclu, vu qu'on est que 2 y'aura pas vraiment de compet, l'organisateur du concours a donc décider de prendre le prix global et de diviser par 2 (donc 250£ chaqu'un)

    il va quand même prendre en compte les votes, mais bon, c'est un peu dommage qu'il n'y est pas vraiment eu de compétition :(


  7. je ne pense pas, car dans les règles du concours il est précisé de fournir une version sans aucune mention de l'auteur du jeu pour le vote publique.

    déjà que certain sont au courant sur le channel #jagware, je voudrais pas fausser les votes %)

    mais je pense que je ne résisterais pas a vous faire une demo à l'ac :D


  8. yeaaaaaaaaah

    finally, I finished my jagcode entry, so ...

    who participated ? :)

     

    I hope we are more than 2 :lol:

     

     

    [French]

    Liiiiiiiiiiiibbreeeeeeeeee

    raaaaaaaaaaah

    j'vais pouvoir glandeeeeeeeer

    ah non merde y'a seaplane :whistling:, pas taper MK !! :lol:

     

    mine de rien, ça fait déja 2 jeux fini sur jag (et 2 gratuits !!)

    je crois que depuis la TI j'ai pas eu autant de projet fini sur une console :blink:

    promis, apres seaplane je fait un comeback :hug:

    (mais pas pour autant que je lache jagware hein :P )


  9. I'm already using this algorithm, but I have "holes" between 2 triangles sometimes, and I think this is because of the imprecision of the increment (24.8 or 16.16 give the same result ...)

    so I thought that using bresenham I would have more accurate results :)


  10. ok, I'm not really good for doing math/algo stuff, so I will try to explain my problem here

    (in english, so prepare for confusing explanation Oo)

     

    I understand how work bresenham line drawing algorithm, but my problem is when you try to apply this algorithm to triangle filling algorithm.

    the basic implementation of triangle filling is to increment Y each step and fill from x1 to x2

    using bresenham with a Y delta larger than a X delta is ok because Y will be incremented each step.

    now my problem is when X delta is larger than the Y delta, because, to know the next X position at the next Y step we must do some kind of loop to compute the X (and doing that for each segment of the triangle depending on deltas ...)

    does anyone get the idea ? :D

    I hope so ...

     

    so, do we must compute the X like that .. (quite boring to code :/) or is there another algorithm/tips ?

×