Skip to content

L
libswscale
Commit 2b0698b4 authored by ramiro's avatar ramiro
Browse files
Options

Factorize x86 fast_bilinear assembly. 


git-svn-id: file:///var/local/repositories/mplayer/trunk/libswscale@29445 b3059339-0415-0410-9bf9-f77b7e298cf2
parent db42dad5
Hide whitespace changes
Inline Side-by-side
Showing with 12 additions and 24 deletions
swscale_template.c
View file @ 2b0698b4
...@@ -2199,6 +2199,14 @@ static inline void RENAME(hScale)(int16_t *dst, int dstW, const uint8_t *src, in ...@@ -2199,6 +2199,14 @@ static inline void RENAME(hScale)(int16_t *dst, int dstW, const uint8_t *src, in
#endif /* HAVE_MMX */ #endif /* HAVE_MMX */
} }
#define FAST_BILINEAR_X86 \
"subl %%edi, %%esi \n\t" /* src[xx+1] - src[xx] */ \
"imull %%ecx, %%esi \n\t" /* (src[xx+1] - src[xx])*xalpha */ \
"shll $16, %%edi \n\t" \
"addl %%edi, %%esi \n\t" /* src[xx+1]*xalpha + src[xx]*(1-xalpha) */ \
"mov %1, %%"REG_D"\n\t" \
"shrl $9, %%esi \n\t" \
static inline void RENAME(hyscale_fast)(SwsContext *c, int16_t *dst, static inline void RENAME(hyscale_fast)(SwsContext *c, int16_t *dst,
int dstWidth, const uint8_t *src, int srcW, int dstWidth, const uint8_t *src, int srcW,
int xInc) int xInc)
...@@ -2335,24 +2343,14 @@ FUNNY_Y_CODE ...@@ -2335,24 +2343,14 @@ FUNNY_Y_CODE
"1: \n\t" "1: \n\t"
"movzbl (%0, %%"REG_d"), %%edi \n\t" //src[xx] "movzbl (%0, %%"REG_d"), %%edi \n\t" //src[xx]
"movzbl 1(%0, %%"REG_d"), %%esi \n\t" //src[xx+1] "movzbl 1(%0, %%"REG_d"), %%esi \n\t" //src[xx+1]
"subl %%edi, %%esi \n\t" //src[xx+1] - src[xx] FAST_BILINEAR_X86
"imull %%ecx, %%esi \n\t" //(src[xx+1] - src[xx])*xalpha
"shll $16, %%edi \n\t"
"addl %%edi, %%esi \n\t" //src[xx+1]*xalpha + src[xx]*(1-xalpha)
"mov %1, %%"REG_D" \n\t"
"shrl $9, %%esi \n\t"
"movw %%si, (%%"REG_D", %%"REG_a", 2) \n\t" "movw %%si, (%%"REG_D", %%"REG_a", 2) \n\t"
"addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF "addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF
"adc %3, %%"REG_d" \n\t" //xx+= xInc>>16 + carry "adc %3, %%"REG_d" \n\t" //xx+= xInc>>16 + carry
"movzbl (%0, %%"REG_d"), %%edi \n\t" //src[xx] "movzbl (%0, %%"REG_d"), %%edi \n\t" //src[xx]
"movzbl 1(%0, %%"REG_d"), %%esi \n\t" //src[xx+1] "movzbl 1(%0, %%"REG_d"), %%esi \n\t" //src[xx+1]
"subl %%edi, %%esi \n\t" //src[xx+1] - src[xx] FAST_BILINEAR_X86
"imull %%ecx, %%esi \n\t" //(src[xx+1] - src[xx])*xalpha
"shll $16, %%edi \n\t"
"addl %%edi, %%esi \n\t" //src[xx+1]*xalpha + src[xx]*(1-xalpha)
"mov %1, %%"REG_D" \n\t"
"shrl $9, %%esi \n\t"
"movw %%si, 2(%%"REG_D", %%"REG_a", 2) \n\t" "movw %%si, 2(%%"REG_D", %%"REG_a", 2) \n\t"
"addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF "addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF
"adc %3, %%"REG_d" \n\t" //xx+= xInc>>16 + carry "adc %3, %%"REG_d" \n\t" //xx+= xInc>>16 + carry
...@@ -2544,22 +2542,12 @@ FUNNY_UV_CODE ...@@ -2544,22 +2542,12 @@ FUNNY_UV_CODE
"mov %0, %%"REG_S" \n\t" "mov %0, %%"REG_S" \n\t"
"movzbl (%%"REG_S", %%"REG_d"), %%edi \n\t" //src[xx] "movzbl (%%"REG_S", %%"REG_d"), %%edi \n\t" //src[xx]
"movzbl 1(%%"REG_S", %%"REG_d"), %%esi \n\t" //src[xx+1] "movzbl 1(%%"REG_S", %%"REG_d"), %%esi \n\t" //src[xx+1]
"subl %%edi, %%esi \n\t" //src[xx+1] - src[xx] FAST_BILINEAR_X86
"imull %%ecx, %%esi \n\t" //(src[xx+1] - src[xx])*xalpha
"shll $16, %%edi \n\t"
"addl %%edi, %%esi \n\t" //src[xx+1]*xalpha + src[xx]*(1-xalpha)
"mov %1, %%"REG_D" \n\t"
"shrl $9, %%esi \n\t"
"movw %%si, (%%"REG_D", %%"REG_a", 2) \n\t" "movw %%si, (%%"REG_D", %%"REG_a", 2) \n\t"
"movzbl (%5, %%"REG_d"), %%edi \n\t" //src[xx] "movzbl (%5, %%"REG_d"), %%edi \n\t" //src[xx]
"movzbl 1(%5, %%"REG_d"), %%esi \n\t" //src[xx+1] "movzbl 1(%5, %%"REG_d"), %%esi \n\t" //src[xx+1]
"subl %%edi, %%esi \n\t" //src[xx+1] - src[xx] FAST_BILINEAR_X86
"imull %%ecx, %%esi \n\t" //(src[xx+1] - src[xx])*xalpha
"shll $16, %%edi \n\t"
"addl %%edi, %%esi \n\t" //src[xx+1]*xalpha + src[xx]*(1-xalpha)
"mov %1, %%"REG_D" \n\t"
"shrl $9, %%esi \n\t"
"movw %%si, "AV_STRINGIFY(VOF)"(%%"REG_D", %%"REG_a", 2) \n\t" "movw %%si, "AV_STRINGIFY(VOF)"(%%"REG_D", %%"REG_a", 2) \n\t"
"addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF "addw %4, %%cx \n\t" //xalpha += xInc&0xFFFF
......
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