I have started coding in 2 libs, either SDL or TinyPTC
(http://www.gaffer.org:8080/tinyptc/), making tests
about their speed.
A test on my PC with a plasma effect at 64048032bpp
and an FPS counter, showed TPTC to be more than twice
faster than SDL (SDL=145FPS, TPTC=355). I removed the
calculation and left only the output just to be sure I
am not making something wrong. Same results…
It’s not the same in every PC. But in most PCs, TPTC
is faster (because of MMX blitting routines they say).
There were some disagreements about that from some
coder friends, but anyways I discovered this thing
acidentally:
If you run an SDL app in a window, then if you move
the mouse pointer upon a button on the programms bar,
suddenly SDL outputs much faster (almost reaching
TPTC) while the bright yellow box with the programm
description appears. Hard to explain, but this means
that SDL can output to the vram much faster but
something (system? OS? A bug?) blocks it from
achieving the best possible…
I have tried with HW_SURFACE or SW_SURFACE, Double
Buffer (SDL_FLIP) in HW_SURFACE, Windowed or
Fullscreen, everytime the same results.
I could compile TPTC under SDL (there is a port) to
have both gfx speed and the rest from SDL that TPTC
misses (keyboard, sound, etc.) but for the moment I
had problems with that…
Or I could find out what’s wrong and how can I make
SDL output the best possible. Because I have seen it
is, but something is going wrong with either my code,
the system or anything…
And I am getting very curious about the matter…
Optimus
p.s. I am sorry about too much text. I hope you can
give me some clue, someone must have an idea around
here. Ask me any questions, I will try to become more
clear and brief…__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com