Commit ab5a1435 authored by Guillaume Poirier's avatar Guillaume Poirier

Move TRANSPOSE8 macro to dsputil_altivec.h.

(was duplicated in libavcodec/ppc/vc1dsp_altivec.c
nd libavcodec/ppc/mpegvideo_altivec.c, and will be
used in upcoming routines for h264_altivec.c)

Originally committed as revision 6621 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent a22b7322
......@@ -67,6 +67,40 @@ void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size,
#define vcii(a,b,c,d) (const vector float){FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d}
#endif
// Transpose 8x8 matrix of 16-bit elements (in-place)
#define TRANSPOSE8(a,b,c,d,e,f,g,h) \
do { \
vector signed short A1, B1, C1, D1, E1, F1, G1, H1; \
vector signed short A2, B2, C2, D2, E2, F2, G2, H2; \
\
A1 = vec_mergeh (a, e); \
B1 = vec_mergel (a, e); \
C1 = vec_mergeh (b, f); \
D1 = vec_mergel (b, f); \
E1 = vec_mergeh (c, g); \
F1 = vec_mergel (c, g); \
G1 = vec_mergeh (d, h); \
H1 = vec_mergel (d, h); \
\
A2 = vec_mergeh (A1, E1); \
B2 = vec_mergel (A1, E1); \
C2 = vec_mergeh (B1, F1); \
D2 = vec_mergel (B1, F1); \
E2 = vec_mergeh (C1, G1); \
F2 = vec_mergel (C1, G1); \
G2 = vec_mergeh (D1, H1); \
H2 = vec_mergel (D1, H1); \
\
a = vec_mergeh (A2, E2); \
b = vec_mergel (A2, E2); \
c = vec_mergeh (B2, F2); \
d = vec_mergel (B2, F2); \
e = vec_mergeh (C2, G2); \
f = vec_mergel (C2, G2); \
g = vec_mergeh (D2, H2); \
h = vec_mergel (D2, H2); \
} while (0)
#endif /* HAVE_ALTIVEC */
#endif /* _DSPUTIL_ALTIVEC_ */
......@@ -52,39 +52,6 @@ do { \
d = vec_mergel(_trans_acl, _trans_bdl); \
} while (0)
#define TRANSPOSE8(a,b,c,d,e,f,g,h) \
do { \
__typeof__(a) _A1, _B1, _C1, _D1, _E1, _F1, _G1, _H1; \
__typeof__(a) _A2, _B2, _C2, _D2, _E2, _F2, _G2, _H2; \
\
_A1 = vec_mergeh (a, e); \
_B1 = vec_mergel (a, e); \
_C1 = vec_mergeh (b, f); \
_D1 = vec_mergel (b, f); \
_E1 = vec_mergeh (c, g); \
_F1 = vec_mergel (c, g); \
_G1 = vec_mergeh (d, h); \
_H1 = vec_mergel (d, h); \
\
_A2 = vec_mergeh (_A1, _E1); \
_B2 = vec_mergel (_A1, _E1); \
_C2 = vec_mergeh (_B1, _F1); \
_D2 = vec_mergel (_B1, _F1); \
_E2 = vec_mergeh (_C1, _G1); \
_F2 = vec_mergel (_C1, _G1); \
_G2 = vec_mergeh (_D1, _H1); \
_H2 = vec_mergel (_D1, _H1); \
\
a = vec_mergeh (_A2, _E2); \
b = vec_mergel (_A2, _E2); \
c = vec_mergeh (_B2, _F2); \
d = vec_mergel (_B2, _F2); \
e = vec_mergeh (_C2, _G2); \
f = vec_mergel (_C2, _G2); \
g = vec_mergeh (_D2, _H2); \
h = vec_mergel (_D2, _H2); \
} while (0)
// Loads a four-byte value (int or float) from the target address
// into every element in the target vector. Only works if the
......
......@@ -26,40 +26,6 @@
#include "dsputil_altivec.h"
// Transpose 8x8 matrix of 16-bit elements. Borrowed from mpegvideo_altivec.c
#define TRANSPOSE8(a,b,c,d,e,f,g,h) \
do { \
vector signed short A1, B1, C1, D1, E1, F1, G1, H1; \
vector signed short A2, B2, C2, D2, E2, F2, G2, H2; \
\
A1 = vec_mergeh (a, e); \
B1 = vec_mergel (a, e); \
C1 = vec_mergeh (b, f); \
D1 = vec_mergel (b, f); \
E1 = vec_mergeh (c, g); \
F1 = vec_mergel (c, g); \
G1 = vec_mergeh (d, h); \
H1 = vec_mergel (d, h); \
\
A2 = vec_mergeh (A1, E1); \
B2 = vec_mergel (A1, E1); \
C2 = vec_mergeh (B1, F1); \
D2 = vec_mergel (B1, F1); \
E2 = vec_mergeh (C1, G1); \
F2 = vec_mergel (C1, G1); \
G2 = vec_mergeh (D1, H1); \
H2 = vec_mergel (D1, H1); \
\
a = vec_mergeh (A2, E2); \
b = vec_mergel (A2, E2); \
c = vec_mergeh (B2, F2); \
d = vec_mergel (B2, F2); \
e = vec_mergeh (C2, G2); \
f = vec_mergel (C2, G2); \
g = vec_mergeh (D2, H2); \
h = vec_mergel (D2, H2); \
} while (0)
// main steps of 8x8 transform
#define STEP8(s0, s1, s2, s3, s4, s5, s6, s7, vec_rnd) \
do { \
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment