liz/rawTherapee
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Simpler interface for gaussian blur, speedup for double precision gaussian blur and speedup for retinex transmission curve

Browse Source
This commit is contained in:
heckflosse
2015-09-16 00:14:58 +02:00
parent 6eba3914f4
commit 5f97c4f15b
8 changed files with 132 additions and 431 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
rtengine/helpersse2.h
View File

@@ -26,15 +26,18 @@ typedef __m128i vint2;
#define LVF(x) _mm_load_ps(&x)
#define LVFU(x) _mm_loadu_ps(&x)
#define STVF(x,y) _mm_store_ps(&x,y)
#define STVFU(x,y) _mm_storeu_ps(&x,y)
#else // there is a bug in gcc 4.7.x when using openmp and aligned memory and -O3
#define LVF(x) _mm_loadu_ps(&x)
#define LVFU(x) _mm_loadu_ps(&x)
#define STVF(x,y) _mm_storeu_ps(&x,y)
#define STVFU(x,y) _mm_storeu_ps(&x,y)
#endif
#else
#define LVF(x) _mm_load_ps(&x)
#define LVFU(x) _mm_loadu_ps(&x)
#define STVF(x,y) _mm_store_ps(&x,y)
#define STVFU(x,y) _mm_storeu_ps(&x,y)
#endif
// Load 8 floats from a and combine a[0],a[2],a[4] and a[6] into a vector of 4 floats