The mmxext implementation is slower than the C version in x86_64. m32 m64 rgb24toyv12_16_200_c: 24942.7 14812.6 rgb24toyv12_16_200_mmxext: 17857.2 ( 1.40x) 17400.4 ( 0.85x) rgb24toyv12_128_60_c: 56892.9 35616.9 rgb24toyv12_128_60_mmxext: 40730.9 ( 1.40x) 39610.4 ( 0.90x) rgb24toyv12_512_16_c: 58402.7 37209.4 rgb24toyv12_512_16_mmxext: 44842.4 ( 1.30x) 41136.2 ( 0.90x) rgb24toyv12_1920_4_c: 54827.4 34737.4 rgb24toyv12_1920_4_mmxext: 51169.9 ( 1.07x) 34818.9 ( 1.00x)
Unified · Split
| @@ -1480,7 +1480,7 @@ static inline void planar2x_mmxext(const uint8_t *src, uint8_t *dst, int srcWidt · 1479 unmodified lines | |||
| 1480 | * others are ignored in the C version. | 1480 | * others are ignored in the C version. |
| 1481 | * FIXME: Write HQ version. | 1481 | * FIXME: Write HQ version. |
| 1482 | */ | 1482 | */ |
| 1483 | #if HAVE_7REGS | 1483 | #if ARCH_X86_32 && HAVE_7REGS |
| 1484 | static inline void rgb24toyv12_mmxext(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, | 1484 | static inline void rgb24toyv12_mmxext(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, |
| 1485 | int width, int height, | 1485 | int width, int height, |
| 1486 | int lumStride, int chromStride, int srcStride, | 1486 | int lumStride, int chromStride, int srcStride, |
| @@ -2257,9 +2257,9 @@ static av_cold void rgb2rgb_init_mmxext(void) · 770 unmodified lines | |||
| 2257 | yuyvtoyuv422 = yuyvtoyuv422_mmxext; | 2257 | yuyvtoyuv422 = yuyvtoyuv422_mmxext; |
| 2258 | 2258 | ||
| 2259 | planar2x = planar2x_mmxext; | 2259 | planar2x = planar2x_mmxext; |
| 2260 | #if HAVE_7REGS | 2260 | #if ARCH_X86_32 && HAVE_7REGS |
| 2261 | ff_rgb24toyv12 = rgb24toyv12_mmxext; | 2261 | ff_rgb24toyv12 = rgb24toyv12_mmxext; |
| 2262 | #endif /* HAVE_7REGS */ | 2262 | #endif /* ARCH_X86_32 && HAVE_7REGS */ |
| 2263 | 2263 | ||
| 2264 | yuyvtoyuv420 = yuyvtoyuv420_mmxext; | 2264 | yuyvtoyuv420 = yuyvtoyuv420_mmxext; |
| 2265 | uyvytoyuv420 = uyvytoyuv420_mmxext; | 2265 | uyvytoyuv420 = uyvytoyuv420_mmxext; |