/external/libvpx/libvpx/vpx_dsp/arm/ |
D | sad4d_neon.c | 17 static INLINE unsigned int horizontal_long_add_16x8(const uint16x8_t vec_lo, in horizontal_long_add_16x8() 18 const uint16x8_t vec_hi) { in horizontal_long_add_16x8() 38 uint16x8_t *vec_sum_ref_lo, in sad_neon_64() 39 uint16x8_t *vec_sum_ref_hi) { in sad_neon_64() 68 uint16x8_t *vec_sum_ref_lo, in sad_neon_32() 69 uint16x8_t *vec_sum_ref_hi) { in sad_neon_32() 87 uint16x8_t vec_sum_ref0_lo = vdupq_n_u16(0); in vpx_sad64x64x4d_neon() 88 uint16x8_t vec_sum_ref0_hi = vdupq_n_u16(0); in vpx_sad64x64x4d_neon() 89 uint16x8_t vec_sum_ref1_lo = vdupq_n_u16(0); in vpx_sad64x64x4d_neon() 90 uint16x8_t vec_sum_ref1_hi = vdupq_n_u16(0); in vpx_sad64x64x4d_neon() [all …]
|
D | sad_neon.c | 23 uint16x8_t q12; in vpx_sad8x16_neon() 57 uint16x8_t q12; in vpx_sad4x4_neon() 88 uint16x8_t q12, q13; in vpx_sad16x8_neon() 119 static INLINE unsigned int horizontal_long_add_16x8(const uint16x8_t vec_lo, in horizontal_long_add_16x8() 120 const uint16x8_t vec_hi) { in horizontal_long_add_16x8() 131 static INLINE unsigned int horizontal_add_16x8(const uint16x8_t vec_16x8) { in horizontal_add_16x8() 142 uint16x8_t vec_accum_lo = vdupq_n_u16(0); in vpx_sad64x64_neon() 143 uint16x8_t vec_accum_hi = vdupq_n_u16(0); in vpx_sad64x64_neon() 178 uint16x8_t vec_accum_lo = vdupq_n_u16(0); in vpx_sad32x32_neon() 179 uint16x8_t vec_accum_hi = vdupq_n_u16(0); in vpx_sad32x32_neon() [all …]
|
D | subtract_neon.c | 29 const uint16x8_t v_diff_lo_00 = vsubl_u8(vget_low_u8(v_src_00), in vpx_subtract_block_neon() 31 const uint16x8_t v_diff_hi_00 = vsubl_u8(vget_high_u8(v_src_00), in vpx_subtract_block_neon() 33 const uint16x8_t v_diff_lo_16 = vsubl_u8(vget_low_u8(v_src_16), in vpx_subtract_block_neon() 35 const uint16x8_t v_diff_hi_16 = vsubl_u8(vget_high_u8(v_src_16), in vpx_subtract_block_neon() 50 const uint16x8_t v_diff_lo = vsubl_u8(vget_low_u8(v_src), in vpx_subtract_block_neon() 52 const uint16x8_t v_diff_hi = vsubl_u8(vget_high_u8(v_src), in vpx_subtract_block_neon() 64 const uint16x8_t v_diff = vsubl_u8(v_src, v_pred); in vpx_subtract_block_neon()
|
D | subpel_variance_neon.c | 44 const uint16x8_t a = vmull_u8(src_0, f0); in var_filter_block2d_bil_w8() 45 const uint16x8_t b = vmlal_u8(a, src_1, f1); in var_filter_block2d_bil_w8() 68 const uint16x8_t a = vmull_u8(vget_low_u8(src_0), f0); in var_filter_block2d_bil_w16() 69 const uint16x8_t b = vmlal_u8(a, vget_low_u8(src_1), f1); in var_filter_block2d_bil_w16() 71 const uint16x8_t c = vmull_u8(vget_high_u8(src_0), f0); in var_filter_block2d_bil_w16() 72 const uint16x8_t d = vmlal_u8(c, vget_high_u8(src_1), f1); in var_filter_block2d_bil_w16()
|
D | intrapred_neon.c | 24 uint16x8_t sum_top; in dc_4x4() 25 uint16x8_t sum_left; in dc_4x4() 43 const uint16x8_t sum = vaddq_u16(sum_left, sum_top); in dc_4x4() 93 uint16x8_t sum_top; in dc_8x8() 94 uint16x8_t sum_left; in dc_8x8() 114 const uint16x8_t sum = vaddq_u16(sum_left, sum_top); in dc_8x8() 164 uint16x8_t sum_top; in dc_16x16() 165 uint16x8_t sum_left; in dc_16x16() 170 const uint16x8_t p0 = vpaddlq_u8(A); // cascading summation of the top in dc_16x16() 179 const uint16x8_t p0 = vpaddlq_u8(L); // cascading summation of the left in dc_16x16() [all …]
|
/external/clang/test/CodeGen/ |
D | arm64_neon_high_half.c | 21 uint16x8_t test_vaddw_high_u8(uint16x8_t lhs, uint8x16_t rhs) { in test_vaddw_high_u8() 26 uint32x4_t test_vaddw_high_u16(uint32x4_t lhs, uint16x8_t rhs) { in test_vaddw_high_u16() 51 uint16x8_t test_vsubw_high_u8(uint16x8_t lhs, uint8x16_t rhs) { in test_vsubw_high_u8() 56 uint32x4_t test_vsubw_high_u16(uint32x4_t lhs, uint16x8_t rhs) { in test_vsubw_high_u16() 81 uint16x8_t test_vabdl_high_u8(uint8x16_t lhs, uint8x16_t rhs) { in test_vabdl_high_u8() 86 uint32x4_t test_vabdl_high_u16(uint16x8_t lhs, uint16x8_t rhs) { in test_vabdl_high_u16() 111 uint16x8_t test_vabal_high_u8(uint16x8_t accum, uint8x16_t lhs, uint8x16_t rhs) { in test_vabal_high_u8() 116 uint32x4_t test_vabal_high_u16(uint32x4_t accum, uint16x8_t lhs, uint16x8_t rhs) { in test_vabal_high_u16() 231 uint16x8_t test_vsubl_high_u8(uint8x16_t lhs, uint8x16_t rhs) { in test_vsubl_high_u8() 236 uint32x4_t test_vsubl_high_u16(uint16x8_t lhs, uint16x8_t rhs) { in test_vsubl_high_u16() [all …]
|
D | aarch64-neon-3v.c | 74 uint16x8_t test_vandq_u16(uint16x8_t a, uint16x8_t b) { in test_vandq_u16() 170 uint16x8_t test_vorrq_u16(uint16x8_t a, uint16x8_t b) { in test_vorrq_u16() 266 uint16x8_t test_veorq_u16(uint16x8_t a, uint16x8_t b) { in test_veorq_u16() 362 uint16x8_t test_vbicq_u16(uint16x8_t a, uint16x8_t b) { in test_vbicq_u16() 458 uint16x8_t test_vornq_u16(uint16x8_t a, uint16x8_t b) { in test_vornq_u16()
|
D | aarch64-neon-2velem.c | 134 uint16x8_t test_vmulq_lane_u16(uint16x8_t a, uint16x4_t v) { in test_vmulq_lane_u16() 176 uint16x4_t test_vmul_laneq_u16(uint16x4_t a, uint16x8_t v) { in test_vmul_laneq_u16() 182 uint16x8_t test_vmulq_laneq_u16(uint16x8_t a, uint16x8_t v) { in test_vmulq_laneq_u16() 524 uint32x4_t test_vmull_high_lane_u16(uint16x8_t a, uint16x4_t v) { in test_vmull_high_lane_u16() 548 uint32x4_t test_vmull_laneq_u16(uint16x4_t a, uint16x8_t v) { in test_vmull_laneq_u16() 572 uint32x4_t test_vmull_high_laneq_u16(uint16x8_t a, uint16x8_t v) { in test_vmull_high_laneq_u16() 941 uint16x8_t test_vmulq_lane_u16_0(uint16x8_t a, uint16x4_t v) { in test_vmulq_lane_u16_0() 983 uint16x4_t test_vmul_laneq_u16_0(uint16x4_t a, uint16x8_t v) { in test_vmul_laneq_u16_0() 989 uint16x8_t test_vmulq_laneq_u16_0(uint16x8_t a, uint16x8_t v) { in test_vmulq_laneq_u16_0() 1295 uint32x4_t test_vmull_high_lane_u16_0(uint16x8_t a, uint16x4_t v) { in test_vmull_high_lane_u16_0() [all …]
|
D | arm_neon_intrinsics.c | 74 uint16x8_t test_vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c) { in test_vabaq_u16() 105 uint16x8_t test_vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c) { in test_vabal_u8() 190 uint16x8_t test_vabdq_u16(uint16x8_t a, uint16x8_t b) { in test_vabdq_u16() 227 uint16x8_t test_vabdl_u8(uint8x8_t a, uint8x8_t b) { in test_vabdl_u8() 385 uint16x8_t test_vaddq_u16(uint16x8_t a, uint16x8_t b) { in test_vaddq_u16() 422 uint8x8_t test_vaddhn_u16(uint16x8_t a, uint16x8_t b) { in test_vaddhn_u16() 459 uint16x8_t test_vaddl_u8(uint8x8_t a, uint8x8_t b) { in test_vaddl_u8() 496 uint16x8_t test_vaddw_u8(uint16x8_t a, uint8x8_t b) { in test_vaddw_u8() 593 uint16x8_t test_vandq_u16(uint16x8_t a, uint16x8_t b) { in test_vandq_u16() 690 uint16x8_t test_vbicq_u16(uint16x8_t a, uint16x8_t b) { in test_vbicq_u16() [all …]
|
D | aarch64-neon-intrinsics.c | 105 uint16x8_t test_vaddq_u16(uint16x8_t v1, uint16x8_t v2) { in test_vaddq_u16() 217 uint16x8_t test_vsubq_u16(uint16x8_t v1, uint16x8_t v2) { in test_vsubq_u16() 302 uint16x8_t test_vmulq_u16(uint16x8_t v1, uint16x8_t v2) { in test_vmulq_u16() 411 uint16x8_t test_vmlaq_u16(uint16x8_t v1, uint16x8_t v2, uint16x8_t v3) { in test_vmlaq_u16() 499 uint16x8_t test_vmlsq_u16(uint16x8_t v1, uint16x8_t v2, uint16x8_t v3) { in test_vmlsq_u16() 629 uint16x8_t test_vabaq_u16(uint16x8_t v1, uint16x8_t v2, uint16x8_t v3) { in test_vabaq_u16() 707 uint16x8_t test_vabdq_u16(uint16x8_t v1, uint16x8_t v2) { in test_vabdq_u16() 810 int16x8_t test_vbslq_s16(uint16x8_t v1, int16x8_t v2, int16x8_t v3) { in test_vbslq_s16() 834 uint16x8_t test_vbslq_u16(uint16x8_t v1, uint16x8_t v2, uint16x8_t v3) { in test_vbslq_u16() 864 poly16x8_t test_vbslq_p16(uint16x8_t v1, poly16x8_t v2, poly16x8_t v3) { in test_vbslq_p16() [all …]
|
D | arm64_vecCmpBr.c | 22 unsigned anyZero128(uint16x8_t a) { in anyZero128() 48 unsigned anyNonZero128(uint16x8_t a) { in anyNonZero128() 74 unsigned allZero128(uint16x8_t a) { in allZero128() 100 unsigned allNonZero128(uint16x8_t a) { in allNonZero128()
|
D | aarch64-neon-perm.c | 68 uint16x8_t test_vuzp1q_u16(uint16x8_t a, uint16x8_t b) { in test_vuzp1q_u16() 194 uint16x8_t test_vuzp2q_u16(uint16x8_t a, uint16x8_t b) { in test_vuzp2q_u16() 320 uint16x8_t test_vzip1q_u16(uint16x8_t a, uint16x8_t b) { in test_vzip1q_u16() 446 uint16x8_t test_vzip2q_u16(uint16x8_t a, uint16x8_t b) { in test_vzip2q_u16() 572 uint16x8_t test_vtrn1q_u16(uint16x8_t a, uint16x8_t b) { in test_vtrn1q_u16() 698 uint16x8_t test_vtrn2q_u16(uint16x8_t a, uint16x8_t b) { in test_vtrn2q_u16() 843 uint16x8x2_t test_vuzpq_u16(uint16x8_t a, uint16x8_t b) { in test_vuzpq_u16() 953 uint16x8x2_t test_vzipq_u16(uint16x8_t a, uint16x8_t b) { in test_vzipq_u16() 1063 uint16x8x2_t test_vtrnq_u16(uint16x8_t a, uint16x8_t b) { in test_vtrnq_u16()
|
D | aarch64-neon-misc.c | 53 uint16x8_t test_vceqzq_s16(int16x8_t a) { in test_vceqzq_s16() 95 uint16x8_t test_vceqzq_u16(uint16x8_t a) { in test_vceqzq_u16() 149 uint16x8_t test_vceqzq_p16(poly16x8_t a) { in test_vceqzq_p16() 197 uint16x8_t test_vcgezq_s16(int16x8_t a) { in test_vcgezq_s16() 269 uint16x8_t test_vclezq_s16(int16x8_t a) { in test_vclezq_s16() 341 uint16x8_t test_vcgtzq_s16(int16x8_t a) { in test_vcgtzq_s16() 413 uint16x8_t test_vcltzq_s16(int16x8_t a) { in test_vcltzq_s16() 545 uint16x8_t test_vrev32q_u16(uint16x8_t a) { in test_vrev32q_u16() 641 uint16x8_t test_vrev64q_u16(uint16x8_t a) { in test_vrev64q_u16() 723 uint16x8_t test_vpaddlq_u8(uint8x16_t a) { in test_vpaddlq_u8() [all …]
|
D | arm64_vqmov.c | 16 uint8x16_t test_vqmovun_high_s16(uint8x8_t Vdlow, uint16x8_t Vn) in test_vqmovun_high_s16() 32 uint16x8_t test_vqmovn_high_u32(uint16x4_t Vdlow, uint32x4_t Vn) in test_vqmovn_high_u32() 56 uint8x16_t test_vqmovn_high_u16(uint8x8_t Vdlow, uint16x8_t Vn) in test_vqmovn_high_u16() 64 uint16x8_t test_vqmovun_high_s32(uint16x4_t Vdlow, uint32x4_t Vn) in test_vqmovun_high_s32()
|
D | struct-matching-constraint.c | 4 typedef __attribute__((neon_vector_type(8))) uint16_t uint16x8_t; typedef 6 void b(uint16x8_t sat, uint16x8_t luma) in b()
|
D | arm64_vshift.c | 83 uint16x8_t test_vqshlq_n_u16(uint16x8_t in) { in test_vqshlq_n_u16() 180 uint16x8_t test_vrshrq_n_u16(uint16x8_t in) { in test_vrshrq_n_u16() 338 uint16x8_t test_vrsraq_n_u16(uint16x8_t acc, uint16x8_t in) { in test_vrsraq_n_u16()
|
D | arm64_vcopy.c | 27 uint16x8_t test_vcopyq_laneq_u16(uint16x8_t a1, uint16x8_t a2) { in test_vcopyq_laneq_u16()
|
D | aarch64-neon-across.c | 57 uint32_t test_vaddlvq_u16(uint16x8_t a) { in test_vaddlvq_u16() 117 uint16_t test_vmaxvq_u16(uint16x8_t a) { in test_vmaxvq_u16() 177 uint16_t test_vminvq_u16(uint16x8_t a) { in test_vminvq_u16() 237 uint16_t test_vaddvq_u16(uint16x8_t a) { in test_vaddvq_u16()
|
D | arm64_vsli.c | 114 uint16x8_t test_vsliq_n_u16(uint16x8_t a1, uint16x8_t a2) { in test_vsliq_n_u16()
|
/external/skia/src/opts/ |
D | SkBlitRow_opts_arm_neon.cpp | 74 uint16x8_t vdst; in S32_D565_Opaque_neon() 110 uint16x8_t vmask_blue, vscale; in S32_D565_Blend_neon() 118 uint16x8_t vdst, vdst_r, vdst_g, vdst_b; in S32_D565_Blend_neon() 119 uint16x8_t vres_r, vres_g, vres_b; in S32_D565_Blend_neon() 490 uint16x8_t vmask_blue; in Color32A_D565_neon() 521 uint16x8_t wide_sr; in Color32A_D565_neon() 522 uint16x8_t wide_sg; in Color32A_D565_neon() 523 uint16x8_t wide_sb; in Color32A_D565_neon() 524 uint16x8_t wide_256_sa; in Color32A_D565_neon() 548 uint16x8_t vdst1, vdst1_r, vdst1_g, vdst1_b; in Color32A_D565_neon() [all …]
|
D | SkColor_opts_neon.h | 21 static inline uint16x8_t SkAlpha255To256_neon8(uint8x8_t alpha) { in SkAlpha255To256_neon8() 25 static inline uint8x8_t SkAlphaMul_neon8(uint8x8_t color, uint16x8_t scale) { in SkAlphaMul_neon8() 29 static inline uint8x8x4_t SkAlphaMulQ_neon8(uint8x8x4_t color, uint16x8_t scale) { in SkAlphaMulQ_neon8() 44 static inline uint8x8x4_t SkPixel16ToPixel32_neon8(uint16x8_t vsrc) { in SkPixel16ToPixel32_neon8() 65 static inline uint16x8_t SkPixel32ToPixel16_neon8(uint8x8x4_t vsrc) { in SkPixel32ToPixel16_neon8() 67 uint16x8_t ret; in SkPixel32ToPixel16_neon8() 79 static inline uint8x8_t SkBlend32_neon8(uint8x8_t src, uint8x8_t dst, uint16x8_t scale) { in SkBlend32_neon8()
|
D | SkBlitMask_opts_arm_neon.cpp | 33 uint16x8_t vmask; in SkBlitLCD16OpaqueRow_neon() 34 uint16x8_t vmaskR, vmaskG, vmaskB; in SkBlitLCD16OpaqueRow_neon() 90 uint16x8_t vcolA; in SkBlitLCD16Row_neon() 101 uint16x8_t vmask; in SkBlitLCD16Row_neon() 102 uint16x8_t vmaskR, vmaskG, vmaskB; in SkBlitLCD16Row_neon() 157 uint16x8_t vdev = vdupq_n_u16(0); in SkRGB16BlitterBlitV_neon() 158 uint16x8_t vmaskq_g16 = vdupq_n_u16(SK_G16_MASK_IN_PLACE); in SkRGB16BlitterBlitV_neon() 159 uint16x8_t vmaskq_ng16 = vdupq_n_u16(~SK_G16_MASK_IN_PLACE); in SkRGB16BlitterBlitV_neon()
|
/external/webp/src/dsp/ |
D | enc_neon.c | 630 const uint16x8_t q_w07 = vld1q_u16(&w[0]); in DistoLoadW() 631 const uint16x8_t q_w8f = vld1q_u16(&w[8]); in DistoLoadW() 728 const uint16x8_t max_coeff_thresh = vdupq_n_u16(MAX_COEFF_THRESH); in CollectHistogram() 738 const uint16x8_t a1 = vreinterpretq_u16_s16(vabsq_s16(a0)); in CollectHistogram() 739 const uint16x8_t b1 = vreinterpretq_u16_s16(vabsq_s16(b0)); in CollectHistogram() 740 const uint16x8_t a2 = vshrq_n_u16(a1, 3); in CollectHistogram() 741 const uint16x8_t b2 = vshrq_n_u16(b1, 3); in CollectHistogram() 742 const uint16x8_t a3 = vminq_u16(a2, max_coeff_thresh); in CollectHistogram() 743 const uint16x8_t b3 = vminq_u16(b2, max_coeff_thresh); in CollectHistogram() 763 uint16x8_t prod = vmull_u8(vget_low_u8(abs_diff), vget_low_u8(abs_diff)); in AccumulateSSE16() [all …]
|
D | dec_neon.c | 1273 const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + 0 * BPS - 1)); in DC4() 1274 const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + 1 * BPS - 1)); in DC4() 1275 const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + 2 * BPS - 1)); in DC4() 1276 const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + 3 * BPS - 1)); in DC4() 1277 const uint16x8_t s0 = vaddq_u16(L0, L1); in DC4() 1278 const uint16x8_t s1 = vaddq_u16(L2, L3); in DC4() 1279 const uint16x8_t s01 = vaddq_u16(s0, s1); in DC4() 1280 const uint16x8_t sum = vaddq_u16(s01, vcombine_u16(p1, p1)); in DC4() 1411 uint16x8_t sum_top; in DC8() 1412 uint16x8_t sum_left; in DC8() [all …]
|
/external/libvpx/libvpx/vp9/encoder/arm/neon/ |
D | vp9_avg_neon.c | 19 static INLINE unsigned int horizontal_add_u16x8(const uint16x8_t v_16x8) { in horizontal_add_u16x8() 30 uint16x8_t v_sum = vaddl_u8(v_s0, v_s1); in vp9_avg_8x8_neon() 56 uint16x8_t vec_sum_lo = vdupq_n_u16(0); in vp9_int_pro_row_neon() 57 uint16x8_t vec_sum_hi = vdupq_n_u16(0); in vp9_int_pro_row_neon() 108 uint16x8_t vec_sum = vdupq_n_u16(0); in vp9_int_pro_col_neon()
|