pandory500/SDL-mirror
2
0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Added a note about aligning SDL_AudioCVT data.

Browse source
This commit is contained in:
Ryan C. Gordon 2017-01-24 00:55:41 -05:00
parent ca2ab5780a
commit e04f23c7ea
1 changed files with 6 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

6
include/SDL_audio.h
View file

@ -185,6 +185,12 @@ typedef void (SDLCALL * SDL_AudioFilter) (struct SDL_AudioCVT * cvt,
/** /**
* A structure to hold a set of audio conversion filters and buffers. * A structure to hold a set of audio conversion filters and buffers.
*
* Note that various parts of the conversion pipeline can take advantage
* of SIMD operations (like SSE2, for example). SDL_AudioCVT doesn't require
* you to pass it aligned data, but can possibly run much faster if you
* set both its (buf) field to a pointer that is aligned to 16 bytes, and its
* (len) field to something that's a multiple of 16, if possible.
*/ */
#ifdef __GNUC__ #ifdef __GNUC__
/* This structure is 84 bytes on 32-bit architectures, make sure GCC doesn't /* This structure is 84 bytes on 32-bit architectures, make sure GCC doesn't