Mercurial > sdl-ios-xcode
view BUGS @ 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 | 9867f3d86e44 |
children | f12379c41042 |
line wrap: on
line source
All: Audio rate conversion is only implemented by multiplying or dividing by a power of two. This is a side-effect of the requirement that the raw audio buffer size be a power of two, and can hopefully be fixed. This means 8 KHz audio converted to 22 KHz ends up being 16 KHz. :-/ When the implementation is writing directly to video memory the mouse cursor doesn't work properly. Applications which do this should use their own mouse cursor and call SDL_ShowCursor(0) to hide the system cursor. Linux: Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed) The keyboard modifiers are not set to the correct state on startup. The AAlib, GGI, and SVGAlib video drivers are not heavily tested. Win32: The MCI driver can't tell if the CD-ROM drive is paused or stopped. The SDL_INIT_EVENTTHREAD flag is not supported on Win32 (Idea: create a separate DirectInput polling thread) The main purpose of this flag is for smooth cursor motion in fullscreen environments. Wide UNICODE character input (Kanji, etc.) is not yet supported. This requires the ToUnicode() API which is only implemented on Windows NT/2000, not on Windows 95/98. Latin-1 keyboard input works fine. Joysticks are only supported under the Win32 MultiMedia API, DirectInput support is not yet implemented. BeOS: BePPC is not supported, apparently not even by Be Inc. SDL_WM_GrabInput() is not implemented. Does anyone know how to do this? SDL_WM_GrabInput() is designed to prevent the user from switching input and mouse focus away from the SDL application. Continuous relative mouse motion is not implemented. Wide UNICODE character input (Kanji, etc.) has not been tested. Latin-1 keyboard input works fine. MacOS: Palette handling isn't implemented in windowed mode yet. Audio hasn't been extensively tested, in particular the locking isn't implemented and mixer routines may not call malloc() or free() because they are called at interrupt time. SDL_WM_GrabInput() is not implemented. Does anyone know how to do this? SDL_WM_GrabInput() is designed to prevent the user from switching input and mouse focus away from the SDL application. Continuous relative mouse motion is not implemented. SDL_AddTimer() and SDL_RemoveTimer() haven't been implemented yet. Not all of the keys are properly recognized on the keyboard. MacOS X: Depth switching for windowed mode isn't implemented yet. Palette handling isn't implemented in windowed mode yet. Command-line arguments dialog is not implemented yet. Fullscreen drawing has some artifacts. Fullscreen OpenGL for the software renderer is broken. Some OpenGL parameters are not accounted for, for example color bits customization. Continuous mouse motion perhaps is not as smooth as it should be. SDL_WM_GrabInput() is implemented, but it "freezes" the hardware cursor in the center of the window/screen. Also, mouse moved events are not generated, and the keyboard cannot be grabbed. MacOS X seems to have a broken pthread_cancel() implementation. Actually, it just only has one cancellation point defined. http://lists.apple.com/archives/darwin-development/2004/Jun/24/killingapthread.001.txt FreeBSD: pthread_cancel() isn't supported by FreeBSD 3.X, so threads don't work on versions of FreeBSD earlier than 4.0. The CD-ROM handling doesn't work completely. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. Solaris: The joystick functions are not implemented yet. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. IRIX: The CD-ROM handling doesn't work completely. The joystick functions are not implemented yet. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. EPOC: Only Crystal version of Epoc/SymbianOS is currently supported. SDL is implemented in static library only. SDL uses static/global variables and that is not allowed in EPOC dll libraries. Sound is not yet supported. Joystick, OpenGL and cdrom is not supported (think about cdrom in a cell phone:-). No console output screen. Printing to stdout do not have any effect. QNX: Fullscreen switch doesn't work (disabled for now). OpenBSD: -= NOT YET SUPPORTED =- This is reported to work, but I haven't verified this. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. OSF/Tru64: -= NOT YET SUPPORTED =- The audio functions are not implemented yet. Joysticks and CD-ROM functions are not implemented yet. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. AIX: -= NOT YET SUPPORTED =- This port has only been tested with AIX 4.3.3 The OpenGL support doesn't work yet. The joystick subsystem isn't implemented yet. Endian detection doesn't work yet - needs a unique CPP symbol. Wide UNICODE character input (Kanji, etc.) is not yet supported. It requires handling of keyboard mapping events and using the XIM input translation extension. I will implement it as requested. Latin-1 keyboard input works fine. (UPDATE 04/06/2004: this bug is now fixed but needs testing) The keyboard modifiers are not set to the correct state on startup. The AIX port was done by Carsten.Griwodz@KOM.tu-darmstadt.de More information on this port is available at: http://www.kom.e-technik.tu-darmstadt.de/~griff/SDL/ AmigaOS: -= NOT YET SUPPORTED =- The OpenGL support isn't implemented yet. SDL_WM_GrabInput() is not implemented. Does anyone know how to do this? SDL_WM_GrabInput() is designed to prevent the user from switching input and mouse focus away from the SDL application. Continuous relative mouse motion is not implemented. The AmigaOS port was done by Gabriele.Greco@galactica.it