mirrors
/
non
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

wscript: Test for __builtin_assume_aligned which is new in GCC 4.7.

This commit is contained in:
Jonathan Moore Liles 2013-08-16 17:34:04 -07:00
parent ca7d95a92d
commit ff41fb9f81
2 changed files with 22 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
nonlib/dsp.C
View File

@ -25,6 +25,12 @@
static const int ALIGNMENT = 16; static const int ALIGNMENT = 16;
#ifdef HAS_BUILTIN_ASSUME_ALIGNED
#define assume_aligned(x) __builtin_assume_aligned(x,ALIGNMENT)
#else
#define assume_aligned(x) (x)
#endif
sample_t * sample_t *
buffer_alloc ( nframes_t size ) buffer_alloc ( nframes_t size )
{ {
@ -38,7 +44,7 @@ buffer_alloc ( nframes_t size )
void void
buffer_apply_gain ( sample_t * __restrict__ buf, nframes_t nframes, float g ) buffer_apply_gain ( sample_t * __restrict__ buf, nframes_t nframes, float g )
{ {
sample_t * buf_ = (sample_t*) __builtin_assume_aligned(buf,ALIGNMENT); sample_t * buf_ = (sample_t*) assume_aligned(buf);
if ( g != 1.0f ) if ( g != 1.0f )
while ( nframes-- ) while ( nframes-- )
@ -56,8 +62,8 @@ buffer_apply_gain_unaligned ( sample_t * __restrict__ buf, nframes_t nframes, fl
void void
buffer_apply_gain_buffer ( sample_t * __restrict__ buf, const sample_t * __restrict__ gainbuf, nframes_t nframes ) buffer_apply_gain_buffer ( sample_t * __restrict__ buf, const sample_t * __restrict__ gainbuf, nframes_t nframes )
{ {
sample_t * buf_ = (sample_t*) __builtin_assume_aligned(buf,ALIGNMENT); sample_t * buf_ = (sample_t*) assume_aligned(buf);
const sample_t * gainbuf_ = (const sample_t*) __builtin_assume_aligned(gainbuf,ALIGNMENT); const sample_t * gainbuf_ = (const sample_t*) assume_aligned(gainbuf);
while ( nframes-- ) while ( nframes-- )
*(buf_++) *= *(gainbuf_++); *(buf_++) *= *(gainbuf_++);
@ -66,9 +72,9 @@ buffer_apply_gain_buffer ( sample_t * __restrict__ buf, const sample_t * __restr
void void
buffer_copy_and_apply_gain_buffer ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, const sample_t * __restrict__ gainbuf, nframes_t nframes ) buffer_copy_and_apply_gain_buffer ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, const sample_t * __restrict__ gainbuf, nframes_t nframes )
{ {
sample_t * dst_ = (sample_t*) __builtin_assume_aligned(dst,ALIGNMENT); sample_t * dst_ = (sample_t*) assume_aligned(dst);
const sample_t * src_ = (const sample_t*) __builtin_assume_aligned(src,ALIGNMENT); const sample_t * src_ = (const sample_t*) assume_aligned(src);
const sample_t * gainbuf_ = (const sample_t*) __builtin_assume_aligned(gainbuf,ALIGNMENT); const sample_t * gainbuf_ = (const sample_t*) assume_aligned(gainbuf);
while ( nframes-- ) while ( nframes-- )
*(dst_++) = *(src_++) * *(gainbuf_++); *(dst_++) = *(src_++) * *(gainbuf_++);
@ -77,8 +83,8 @@ buffer_copy_and_apply_gain_buffer ( sample_t * __restrict__ dst, const sample_t
void void
buffer_mix ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, nframes_t nframes ) buffer_mix ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, nframes_t nframes )
{ {
sample_t * dst_ = (sample_t*) __builtin_assume_aligned(dst,ALIGNMENT); sample_t * dst_ = (sample_t*) assume_aligned(dst);
const sample_t * src_ = (const sample_t*) __builtin_assume_aligned(src,ALIGNMENT); const sample_t * src_ = (const sample_t*) assume_aligned(src);
while ( nframes-- ) while ( nframes-- )
*(dst_++) += *(src_++); *(dst_++) += *(src_++);
@ -87,8 +93,8 @@ buffer_mix ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, nfr
void void
buffer_mix_with_gain ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, nframes_t nframes, float g ) buffer_mix_with_gain ( sample_t * __restrict__ dst, const sample_t * __restrict__ src, nframes_t nframes, float g )
{ {
sample_t * dst_ = (sample_t*) __builtin_assume_aligned(dst,ALIGNMENT); sample_t * dst_ = (sample_t*) assume_aligned(dst);
const sample_t * src_ = (const sample_t*) __builtin_assume_aligned(src,ALIGNMENT); const sample_t * src_ = (const sample_t*) assume_aligned(src);
while ( nframes-- ) while ( nframes-- )
*(dst_++) += *(src_++) * g; *(dst_++) += *(src_++) * g;
@ -152,7 +158,7 @@ buffer_is_digital_black ( sample_t *buf, nframes_t nframes )
float float
buffer_get_peak ( const sample_t * __restrict__ buf, nframes_t nframes ) buffer_get_peak ( const sample_t * __restrict__ buf, nframes_t nframes )
{ {
const sample_t * buf_ = (const sample_t*) __builtin_assume_aligned(buf,ALIGNMENT); const sample_t * buf_ = (const sample_t*) assume_aligned(buf);
float p = 0.0f; float p = 0.0f;
@ -191,7 +197,7 @@ Value_Smoothing_Filter::sample_rate ( nframes_t n )
bool bool
Value_Smoothing_Filter::apply( sample_t * __restrict__ dst, nframes_t nframes, float gt ) Value_Smoothing_Filter::apply( sample_t * __restrict__ dst, nframes_t nframes, float gt )
{ {
sample_t * dst_ = (sample_t*) __builtin_assume_aligned(dst,ALIGNMENT); sample_t * dst_ = (sample_t*) assume_aligned(dst);
const float a = 0.07f; const float a = 0.07f;
const float b = 1 + a; const float b = 1 + a;

4
wscript
View File

@ -63,6 +63,10 @@ def configure(conf):
conf.check_cfg(package='liblo', uselib_store='LIBLO',args="--cflags --libs", conf.check_cfg(package='liblo', uselib_store='LIBLO',args="--cflags --libs",
atleast_version='0.26', mandatory=True) atleast_version='0.26', mandatory=True)
conf.check_cc(msg='Checking for compiler pointer alignment hints',
uselib_store='HAS_BUILTIN_ASSUME_ALIGNED',
fragment='int main ( char**argv, int argc ) { const char *s = (const char*)__builtin_assume_aligned( 0, 16 ); return 0; }',
execute=False, mandatory=False)
### ###
for i in common: for i in common: