Noxid
Reaction score
0

Profile posts Postings Media Albums About

  • Or have you discovered anything new? Most of the particle things online have very little trouble with rendering, and can do up to 10,000 before things get sloppy. huh. Maybe we could try seeing how to get the particles calculated on the gpu itself. That way there will be no bus restriction on particles.

    Also, after a bit more thunkin' the calc thing wouldn't save THAT much memory.

    Also, try running without drawing - where does that cap at?
    So I learned we can use almost no ram (5 bytes per particle) by using calculus. It would almost certainly be faster. I think. It still wouldn't help the rendertime.
    For that I say we opt for a combo of alternating cycle and bltfast.
    Maybe I can do math though.
    we could skip frames then - evenly numbered particles draw one step, odd particles draw the next.
    My only idea so far isn't too good.
    Basically, we draw all the particles to a surface during the loop phase, and then render that surface afterwards.
    holy shit - did you STAY UP to eight o'clock!?
    what do they do exactly - just slow, or...
    It's broken in either the main loop or the collision code - "Access Viloation when writing to 0000000". Ideas?
    aannnnnnnnnnd fix't. I wrote down something wrong.
    The loop successfully doesn't do anything with no particles!
    YAYYYY!!!!!
    just warning you that the loop calls to 400000 instead of the collision code.
    :p

    I'll fix the hex
    also, there is some code still left from undead core that jumps weirdly into the first function.
    hu?
    you sure you typed in the offsets right?
    two of the functions aint snugtight
    Hey, I'm sorry for asking, but how do I upload pictures directly from the computer while replying to a thread?
    Oh.
    Cpt. Fabulous is back...
    Oh well.
    Maybe we can start over and be friends...
    Or, you know, maybe not.
  • Loading…
  • Loading…
  • Loading…
  • Loading…
Top