Lines Matching refs:taps
46 const uint8x8_t* const taps) { in SumOnePassTaps() argument
50 sum = vmull_u8(src[0], taps[0]); in SumOnePassTaps()
52 sum = vmlsl_u8(sum, src[1], taps[1]); in SumOnePassTaps()
53 sum = vmlal_u8(sum, src[2], taps[2]); in SumOnePassTaps()
54 sum = vmlal_u8(sum, src[3], taps[3]); in SumOnePassTaps()
55 sum = vmlsl_u8(sum, src[4], taps[4]); in SumOnePassTaps()
56 sum = vmlal_u8(sum, src[5], taps[5]); in SumOnePassTaps()
60 sum = vmull_u8(src[1], taps[1]); in SumOnePassTaps()
61 sum = vmlsl_u8(sum, src[0], taps[0]); in SumOnePassTaps()
62 sum = vmlal_u8(sum, src[2], taps[2]); in SumOnePassTaps()
63 sum = vmlal_u8(sum, src[3], taps[3]); in SumOnePassTaps()
64 sum = vmlal_u8(sum, src[4], taps[4]); in SumOnePassTaps()
65 sum = vmlsl_u8(sum, src[5], taps[5]); in SumOnePassTaps()
68 sum = vmull_u8(src[0], taps[0]); in SumOnePassTaps()
69 sum = vmlal_u8(sum, src[1], taps[1]); in SumOnePassTaps()
70 sum = vmlal_u8(sum, src[2], taps[2]); in SumOnePassTaps()
71 sum = vmlal_u8(sum, src[3], taps[3]); in SumOnePassTaps()
72 sum = vmlal_u8(sum, src[4], taps[4]); in SumOnePassTaps()
73 sum = vmlal_u8(sum, src[5], taps[5]); in SumOnePassTaps()
76 sum = vmull_u8(src[1], taps[1]); in SumOnePassTaps()
77 sum = vmlsl_u8(sum, src[0], taps[0]); in SumOnePassTaps()
78 sum = vmlsl_u8(sum, src[2], taps[2]); in SumOnePassTaps()
79 sum = vmlal_u8(sum, src[3], taps[3]); in SumOnePassTaps()
80 sum = vmlal_u8(sum, src[4], taps[4]); in SumOnePassTaps()
81 sum = vmlsl_u8(sum, src[5], taps[5]); in SumOnePassTaps()
82 sum = vmlal_u8(sum, src[6], taps[6]); in SumOnePassTaps()
83 sum = vmlsl_u8(sum, src[7], taps[7]); in SumOnePassTaps()
86 sum = vmull_u8(src[0], taps[0]); in SumOnePassTaps()
87 sum = vmlal_u8(sum, src[1], taps[1]); in SumOnePassTaps()
90 sum = vmull_u8(src[1], taps[1]); in SumOnePassTaps()
91 sum = vmlsl_u8(sum, src[0], taps[0]); in SumOnePassTaps()
92 sum = vmlal_u8(sum, src[2], taps[2]); in SumOnePassTaps()
93 sum = vmlsl_u8(sum, src[3], taps[3]); in SumOnePassTaps()
96 sum = vmull_u8(src[0], taps[0]); in SumOnePassTaps()
97 sum = vmlal_u8(sum, src[1], taps[1]); in SumOnePassTaps()
98 sum = vmlal_u8(sum, src[2], taps[2]); in SumOnePassTaps()
99 sum = vmlal_u8(sum, src[3], taps[3]); in SumOnePassTaps()
383 const int16x8_t taps) { in Sum2DVerticalTaps4() argument
384 const int16x4_t taps_lo = vget_low_s16(taps); in Sum2DVerticalTaps4()
385 const int16x4_t taps_hi = vget_high_s16(taps); in Sum2DVerticalTaps4()
422 const int16x8_t taps) { in SimpleSum2DVerticalTaps() argument
423 const int16x4_t taps_lo = vget_low_s16(taps); in SimpleSum2DVerticalTaps()
424 const int16x4_t taps_hi = vget_high_s16(taps); in SimpleSum2DVerticalTaps()
489 const int height, const int16x8_t taps) { in Filter2DVerticalWidth8AndUp() argument
528 SimpleSum2DVerticalTaps<num_taps, is_compound>(srcs + 0, taps); in Filter2DVerticalWidth8AndUp()
530 SimpleSum2DVerticalTaps<num_taps, is_compound>(srcs + 1, taps); in Filter2DVerticalWidth8AndUp()
565 const int16x8_t taps) { in Filter2DVerticalWidth4() argument
596 SimpleSum2DVerticalTaps<num_taps, is_compound>(srcs, taps); in Filter2DVerticalWidth4()
631 const int16x8_t taps) { in Filter2DVerticalWidth2() argument
670 SimpleSum2DVerticalTaps<num_taps, /*is_compound=*/false>(srcs, taps); in Filter2DVerticalWidth2()
743 const int width, const int height, const int16x8_t taps, in Filter2DVertical() argument
748 intermediate_result, dest, pred_stride, width, height, taps); in Filter2DVertical()
751 pred_stride, height, taps); in Filter2DVertical()
755 pred_stride, height, taps); in Filter2DVertical()
787 const int16x8_t taps = vmovl_s8( in Convolve2D_NEON() local
790 Filter2DVertical<8>(intermediate_result, width, height, taps, prediction, in Convolve2D_NEON()
793 Filter2DVertical<6>(intermediate_result, width, height, taps, prediction, in Convolve2D_NEON()
796 Filter2DVertical<4>(intermediate_result, width, height, taps, prediction, in Convolve2D_NEON()
799 Filter2DVertical<2>(intermediate_result, width, height, taps, prediction, in Convolve2D_NEON()
866 const uint8x8_t taps[2] = {VQTbl1U8(filter_taps0, filter_indices), in ConvolveKernelHorizontal2Tap() local
881 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/3>(src, taps), in ConvolveKernelHorizontal2Tap()
906 const uint8x8_t taps[2] = {VQTbl1U8(filter_taps0, filter_indices), in ConvolveKernelHorizontal2Tap() local
921 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/3>(src, taps), in ConvolveKernelHorizontal2Tap()
968 const uint8x8_t taps[4] = {VQTbl1U8(filter_taps0, filter_indices), in ConvolveKernelHorizontalPositive4Tap() local
990 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/5>(src, taps), in ConvolveKernelHorizontalPositive4Tap()
1037 const uint8x8_t taps[4] = {VQTbl1U8(filter_taps0, filter_indices), in ConvolveKernelHorizontalSigned4Tap() local
1063 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/4>(src, taps), in ConvolveKernelHorizontalSigned4Tap()
1128 uint8x8_t taps[6]; in ConvolveKernelHorizontalSigned6Tap() local
1130 taps[i] = VQTbl1U8(filter_taps[i], filter_indices); in ConvolveKernelHorizontalSigned6Tap()
1143 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/0>(src, taps), in ConvolveKernelHorizontalSigned6Tap()
1189 uint8x8_t taps[4]; in ConvolveKernelHorizontalMixed6Tap() local
1225 taps[i] = VQTbl1U8(positive_filter_taps[i], filter_indices); in ConvolveKernelHorizontalMixed6Tap()
1240 sum = vmlal_u8(sum, taps[0], vtbl3_u8(src_vals, src_lookup[1])); in ConvolveKernelHorizontalMixed6Tap()
1241 sum = vmlal_u8(sum, taps[1], vtbl3_u8(src_vals, src_lookup[2])); in ConvolveKernelHorizontalMixed6Tap()
1242 sum = vmlal_u8(sum, taps[2], vtbl3_u8(src_vals, src_lookup[3])); in ConvolveKernelHorizontalMixed6Tap()
1243 sum = vmlal_u8(sum, taps[3], vtbl3_u8(src_vals, src_lookup[4])); in ConvolveKernelHorizontalMixed6Tap()
1282 uint8x8_t taps[8]; in ConvolveKernelHorizontalSigned8Tap() local
1312 taps[i] = VQTbl1U8(filter_taps[i], filter_indices); in ConvolveKernelHorizontalSigned8Tap()
1327 vrshrq_n_s16(SumOnePassTaps</*filter_index=*/2>(src, taps), in ConvolveKernelHorizontalSigned8Tap()
1747 const uint8x8_t* const taps) { in FilterVertical() argument
1788 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical()
1819 const int height, const uint8x8_t* const taps) { in FilterVertical4xH() argument
1841 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical4xH()
1879 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical4xH()
1924 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical4xH()
1976 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical4xH()
2006 const int height, const uint8x8_t* const taps) { in FilterVertical2xH() argument
2032 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical2xH()
2074 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical2xH()
2124 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical2xH()
2181 SumOnePassTaps<filter_index, negative_outside_taps>(srcs, taps); in FilterVertical2xH()
2225 uint8x8_t taps[8]; in ConvolveVertical_NEON() local
2227 taps[k] = in ConvolveVertical_NEON()
2234 taps + 1); in ConvolveVertical_NEON()
2237 taps + 1); in ConvolveVertical_NEON()
2240 taps + 1); in ConvolveVertical_NEON()
2246 taps + 1); in ConvolveVertical_NEON()
2249 taps + 1); in ConvolveVertical_NEON()
2252 taps + 1); in ConvolveVertical_NEON()
2260 src, src_stride, dest, dest_stride, height, taps + 1); in ConvolveVertical_NEON()
2264 src, src_stride, dest, dest_stride, height, taps + 1); in ConvolveVertical_NEON()
2267 src, src_stride, dest, dest_stride, width, height, taps + 1); in ConvolveVertical_NEON()
2271 FilterVertical2xH<2>(src, src_stride, dest, dest_stride, height, taps); in ConvolveVertical_NEON()
2273 FilterVertical4xH<2>(src, src_stride, dest, dest_stride, height, taps); in ConvolveVertical_NEON()
2276 taps); in ConvolveVertical_NEON()
2281 taps + 3); in ConvolveVertical_NEON()
2284 taps + 3); in ConvolveVertical_NEON()
2287 taps + 3); in ConvolveVertical_NEON()
2293 taps + 2); in ConvolveVertical_NEON()
2296 taps + 2); in ConvolveVertical_NEON()
2299 taps + 2); in ConvolveVertical_NEON()
2316 taps + 2); in ConvolveVertical_NEON()
2319 taps + 2); in ConvolveVertical_NEON()
2322 taps + 2); in ConvolveVertical_NEON()
2397 uint8x8_t taps[8]; in ConvolveCompoundVertical_NEON() local
2399 taps[k] = in ConvolveCompoundVertical_NEON()
2406 height, taps + 1); in ConvolveCompoundVertical_NEON()
2409 width, height, taps + 1); in ConvolveCompoundVertical_NEON()
2415 height, taps + 1); in ConvolveCompoundVertical_NEON()
2418 width, height, taps + 1); in ConvolveCompoundVertical_NEON()
2426 4, height, taps + 1); in ConvolveCompoundVertical_NEON()
2429 src, src_stride, dest, width, width, height, taps + 1); in ConvolveCompoundVertical_NEON()
2434 height, taps); in ConvolveCompoundVertical_NEON()
2437 width, height, taps); in ConvolveCompoundVertical_NEON()
2442 height, taps + 3); in ConvolveCompoundVertical_NEON()
2445 width, height, taps + 3); in ConvolveCompoundVertical_NEON()
2450 height, taps + 2); in ConvolveCompoundVertical_NEON()
2453 width, height, taps + 2); in ConvolveCompoundVertical_NEON()
2470 height, taps + 2); in ConvolveCompoundVertical_NEON()
2473 width, height, taps + 2); in ConvolveCompoundVertical_NEON()
2496 const int width, const int height, const int16x8_t taps, in Compound2DVertical() argument
2501 intermediate_result, dest, width, height, taps); in Compound2DVertical()
2504 intermediate_result, dest, width, width, height, taps); in Compound2DVertical()
2541 const int16x8_t taps = vmovl_s8( in ConvolveCompound2D_NEON() local
2544 Compound2DVertical<8>(intermediate_result, width, height, taps, prediction); in ConvolveCompound2D_NEON()
2546 Compound2DVertical<6>(intermediate_result, width, height, taps, prediction); in ConvolveCompound2D_NEON()
2548 Compound2DVertical<4>(intermediate_result, width, height, taps, prediction); in ConvolveCompound2D_NEON()
2550 Compound2DVertical<2>(intermediate_result, width, height, taps, prediction); in ConvolveCompound2D_NEON()