Mercurial > sdl-ios-xcode
view README.wscons @ 1542:a8bf1aa21020
Fixed bug #15
SDL_blit_A.mmx-speed.patch.txt --
Speed improvements and a bugfix for the current GCC inline mmx
asm code:
- Changed some ops and removed some resulting useless ones.
- Added some instruction parallelism (some gain)
The resulting speed on my Xeon improved upto 35% depending on
the function (measured in fps).
- Fixed a bug where BlitRGBtoRGBSurfaceAlphaMMX() was
setting the alpha component on the destination surfaces (to
opaque-alpha) even when the surface had none.
SDL_blit_A.mmx-msvc.patch.txt --
MSVC mmx intrinsics version of the same GCC asm code.
MSVC compiler tries to parallelize the code and to avoid
register stalls, but does not always do a very good job.
Per-surface blending MSVC functions run quite a bit faster
than their pure-asm counterparts (upto 55% faster for 16bit
ones), but the per-pixel blending runs somewhat slower than asm.
- BlitRGBtoRGBSurfaceAlphaMMX and BlitRGBtoRGBPixelAlphaMMX (and all
variants) can now also handle formats other than (A)RGB8888. Formats
like RGBA8888 and some quite exotic ones are allowed -- like
RAGB8888, or actually anything having channels aligned on 8bit
boundary and full 8bit alpha (for per-pixel alpha blending).
The performance cost of this change is virtually 0 for per-surface
alpha blending (no extra ops inside the loop) and a single non-MMX
op inside the loop for per-pixel blending. In testing, the per-pixel
alpha blending takes a ~2% performance hit, but it still runs much
faster than the current code in CVS. If necessary, a separate function
with this functionality can be made.
This code requires Processor Pack for VC6.
author | Sam Lantinga <slouken@libsdl.org> |
---|---|
date | Wed, 15 Mar 2006 15:39:29 +0000 |
parents | 19d8949b4584 |
children |
line wrap: on
line source
============================================================================== Using the Simple DirectMedia Layer with OpenBSD/wscons ============================================================================== The wscons SDL driver can be used to run SDL programs on OpenBSD without running X. So far, the driver only runs on the Sharp Zaurus, but the driver is written to be easily extended for other machines. The main missing pieces are blitting routines for anything but 16 bit displays, and keycode maps for other keyboards. Also, there is no support for hardware palettes. There is currently no mouse support. To compile SDL with support for wscons, use the "--enable-video-wscons" option when running configure. I used the following command line: ./configure --disable-oss --disable-ltdl --enable-pthread-sem \ --disable-esd --disable-arts --disable-video-aalib \ --enable-openbsdaudio --enable-video-wscons \ --prefix=/usr/local --sysconfdir=/etc Setting the console device to use ================================= When starting an SDL program on a wscons console, the driver uses the current virtual terminal (usually /dev/ttyC0). To force the driver to use a specific terminal device, set the environment variable SDL_WSCONSDEV: bash$ SDL_WSCONSDEV=/dev/ttyC1 ./some-sdl-program This is especially useful when starting an SDL program from a remote login prompt (which is great for development). If you do this, and want to use keyboard input, you should avoid having some other program reading from the used virtual console (i.e., do not have a getty running). Rotating the display ==================== The display can be rotated by the wscons SDL driver. This is useful for the Sharp Zaurus, since the display hardware is wired so that it is correctly rotated only when the display is folded into "PDA mode." When using the Zaurus in "normal," or "keyboard" mode, the hardware screen is rotated 90 degrees anti-clockwise. To let the wscons SDL driver rotate the screen, set the environment variable SDL_VIDEO_WSCONS_ROTATION to "CW", "CCW", or "UD", for clockwise, counter clockwise, and upside-down rotation respectively. "CW" makes the screen appear correct on a Sharp Zaurus SL-C3100. When using rotation in the driver, a "shadow" frame buffer is used to hold the intermediary display, before blitting it to the actual hardware frame buffer. This slows down performance a bit. For completeness, the rotation "NONE" can be specified to use a shadow frame buffer without actually rotating. Unsetting SDL_VIDEO_WSCONS_ROTATION, or setting it to '' turns off the shadow frame buffer for maximum performance. Running MAME ============ Since my main motivation for writing the driver was playing MAME on the Zaurus, I'll give a few hints: XMame compiles just fine under OpenBSD. I'm not sure this is strictly necessary, but set MY_CPU = arm in makefile.unix, and CFLAGS.arm = -DLSB_FIRST -DALIGN_INTS -DALIGN_SHORTS in src/unix/unix.max to be sure. The latest XMame (0.101 at this writing) is a very large program. Either tinker with the make files to compile a version without support for all drivers, or, get an older version of XMame. My recommendation would be 0.37b16. When running MAME, DO NOT SET SDL_VIDEO_WSCONS_ROTATION! Performace is MUCH better without this, and it is COMPLETELY UNNECESSARY, since MAME can rotate the picture itself while drawing, and does so MUCH FASTER. Use the Xmame command line option "-ror" to rotate the picture to the right. Acknowledgments =============== I studied the wsfb driver for XFree86/Xorg quite a bit before writing this, so there ought to be some similarities. -- Staffan Ulfberg <staffan@ulfberg.se>