Actually, in Direct3D "full screen mode" is different from a borderless
window covering the whole screen, even when your desktop is set to the same
resolution/bit depth as the game... although it's up to the video driver how
to handle either case, it's pretty normal that in windowed mode, the 3D card
renders to a single offscreen buffer that is copied to the screen with a
sprite or bitblt, while fullscreen mode will use 2 or 3 buffers and switch
between them with a nearly instantaneous page flip (although some FSAA
methods involve using a downsampling sprite copy instead of a page flip).
And of course in full screen the game can run in a different resolution, bit
depth, and refresh rate from the desktop. It is true that games typicaly put
up a big window when running in fullscreen mode, but that's just to make
sure you capture mouse events and such. I'm curious as to why people would
get a speed _increase_ when running in windowed mode with this game--I would
normally expect the opposite. My only hypotheses are: a) perhaps the
fullscreen mode runs in double-buffer mode syncing to the vblank, which
tends to mean that if you can't hit 60 fps, you drop to 30, and if you can't
hit 30, you drop to 20, instead of running at 57 fps or 27 fps respectively,
b) perhaps the game turns on more graphics features by default when in full
screen, such as anisotropic or fsaa which slow it down, or c) perhaps it
sets a lower refresh rate in fullscreen mode, which can move the "sweet
spot" frame rates around when using double buffering, or d) either a driver
bug or a bug in the game, or the game using vertex buffers or hardware t&l
improperly or something like that--there are ways of using Direct3D that are
actually _slower_ with hardware t&l enabled, but one would hope this game
wouldn't be doing those things. Another thought, perhaps they're "retro"
and locking the primary buffer and drawing their overlays on it with 2D
primitives, which I could see as being faster in windowed mode. Maybe we
could bet on the different theories and raffle off a stuffed zebra or
something.
-Dave
> > > > Erm...yeah, that's exactly what I was getting at - why should a game
run
> > > > better in a window if it's the same res on the same graphics card?
> > Doesn't
> > > > make any sense.
> > > uh, guys, full-screen is a borderless window covering the whole
> > > screen.
> > Of course - what's your point?
> that running a game in a window the size of the desktop, and in
> fullscreen doesn't make a difference?
> --
> Gunnar
> #31 SUCKS#015 Tupperware MC#002 DoD#0x1B DoDRT#003 DoD:CT#4,8 Kibo: 2
> to err is human -- to forgive is bovine.