1. Ongoing Graphics Woe...
Well I tried a generic bitmap-drawing procedure that looks like this:
procedure my_display_image(sequence l, sequence b)
atom offx, offy
offx = l[1] - 1
offy = l[2] - 1
for x = 1 to length(b) do
for y = 1 to length(b[x]) do
if b[x][y] != 0 then
pixel(b[x][y],{l[1]+x,l[2]+y})
end if
end for
end for
end procedure
But it was too slow. I still need to write a routine that does this
quickly in 1024x768x256 graphics mode. Anybody have any ideas on how to
do this?
John DeHope
2. Re: Ongoing Graphics Woe...
> But it was too slow. I still need to write a routine that does this
> quickly in 1024x768x256 graphics mode. Anybody have any ideas on how to
> do this?
You should not use pixel...
The video memory of that video mode is also #A0000, however it is
not linear anymore.. (problem).. but with a asm-drawing routines that
use bank wisely( unlike Euporia's pixel)... that is draw everything
in memory and then copy the total to the screen (you want to do this
anyway...against flickering) when copying it draws the first section
and then it banks to the next.. (only once..not for EVEYY pixel in an
other section) then it draws the next section...etc
The next time you draw it starts at the section it has visited the
last time (no backing required thus) and steps backwardly through the
sections...
It's the fastest method i know...however some talked about DMA-pixel
routines..i dunno if they work with SVGA..i dunno anything about DMA
(except that it means direct memory access)
Hope this helps..(and that some ASM-programmer want to code
this...if so.. please like the structure of the extended-memory-copy
by Micheal Bolin...it has the fastest possible interface with the
command buffers)
Ralf
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
~~~~>> Ralf Nieuwenhuijsen <<~~~~
~~~>> nieuwen at xs4all.nl <<~~~~
