/external/libvpx/libvpx/vp8/encoder/arm/neon/ |
D | denoising_neon.c | 61 const uint8x16_t v_level1_adjustment = vmovq_n_u8( in vp8_denoiser_filter_neon() 63 const uint8x16_t v_delta_level_1_and_2 = vdupq_n_u8(1); in vp8_denoiser_filter_neon() 64 const uint8x16_t v_delta_level_2_and_3 = vdupq_n_u8(2); in vp8_denoiser_filter_neon() 65 const uint8x16_t v_level1_threshold = vmovq_n_u8(4 + shift_inc); in vp8_denoiser_filter_neon() 66 const uint8x16_t v_level2_threshold = vdupq_n_u8(8); in vp8_denoiser_filter_neon() 67 const uint8x16_t v_level3_threshold = vdupq_n_u8(16); in vp8_denoiser_filter_neon() 74 const uint8x16_t v_sig = vld1q_u8(sig); in vp8_denoiser_filter_neon() 75 const uint8x16_t v_mc_running_avg_y = vld1q_u8(mc_running_avg_y); in vp8_denoiser_filter_neon() 78 const uint8x16_t v_abs_diff = vabdq_u8(v_sig, v_mc_running_avg_y); in vp8_denoiser_filter_neon() 79 const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg_y); in vp8_denoiser_filter_neon() [all …]
|
/external/webp/src/dsp/ |
D | dec_neon.c | 67 uint8x16_t* const p1, uint8x16_t* const p0, in Load4x16() 68 uint8x16_t* const q0, uint8x16_t* const q1) { in Load4x16() 87 uint8x16_t* const p1, uint8x16_t* const p0, in Load4x16() 88 uint8x16_t* const q0, uint8x16_t* const q1) { in Load4x16() 130 uint8x16_t* const p3, uint8x16_t* const p2, in Load8x16() 131 uint8x16_t* const p1, uint8x16_t* const p0, in Load8x16() 132 uint8x16_t* const q0, uint8x16_t* const q1, in Load8x16() 133 uint8x16_t* const q2, uint8x16_t* const q3) { in Load8x16() 139 uint8x16_t* const p1, uint8x16_t* const p0, in Load16x4() 140 uint8x16_t* const q0, uint8x16_t* const q1) { in Load16x4() [all …]
|
D | lossless_neon.c | 35 const uint8x16_t tmp = pixel.val[0]; in ConvertBGRAToRGBA() 275 static WEBP_INLINE uint8x16_t DoGreenShuffle(const uint8x16_t argb, in DoGreenShuffle() 276 const uint8x16_t shuffle) { in DoGreenShuffle() 284 static WEBP_INLINE uint8x16_t DoGreenShuffle(const uint8x16_t argb, in DoGreenShuffle() 294 const uint8x16_t shuffle = vld1q_u8(kGreenShuffle); in SubtractGreenFromBlueAndRed() 299 const uint8x16_t argb = vld1q_u8((uint8_t*)argb_data); in SubtractGreenFromBlueAndRed() 300 const uint8x16_t greens = DoGreenShuffle(argb, shuffle); in SubtractGreenFromBlueAndRed() 310 const uint8x16_t shuffle = vld1q_u8(kGreenShuffle); in AddGreenToBlueAndRed() 315 const uint8x16_t argb = vld1q_u8((uint8_t*)argb_data); in AddGreenToBlueAndRed() 316 const uint8x16_t greens = DoGreenShuffle(argb, shuffle); in AddGreenToBlueAndRed()
|
/external/clang/test/CodeGen/ |
D | aarch64-neon-tbl.c | 167 uint8x8_t test_vqtbl1_u8(uint8x16_t a, uint8x8_t b) { in test_vqtbl1_u8() 209 uint8x16_t test_vqtbl1q_u8(uint8x16_t a, uint8x16_t b) { in test_vqtbl1q_u8() 215 uint8x16_t test_vqtbl2q_u8(uint8x16x2_t a, uint8x16_t b) { in test_vqtbl2q_u8() 221 uint8x16_t test_vqtbl3q_u8(uint8x16x3_t a, uint8x16_t b) { in test_vqtbl3q_u8() 227 uint8x16_t test_vqtbl4q_u8(uint8x16x4_t a, uint8x16_t b) { in test_vqtbl4q_u8() 265 uint8x8_t test_vqtbx1_u8(uint8x8_t a, uint8x16_t b, uint8x8_t c) { in test_vqtbx1_u8() 289 uint8x16_t test_vqtbx1q_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) { in test_vqtbx1q_u8() 295 uint8x16_t test_vqtbx2q_u8(uint8x16_t a, uint8x16x2_t b, uint8x16_t c) { in test_vqtbx2q_u8() 301 uint8x16_t test_vqtbx3q_u8(uint8x16_t a, uint8x16x3_t b, uint8x16_t c) { in test_vqtbx3q_u8() 307 uint8x16_t test_vqtbx4q_u8(uint8x16_t a, uint8x16x4_t b, uint8x16_t c) { in test_vqtbx4q_u8() [all …]
|
D | arm64_crypto.c | 6 uint8x16_t test_aese(uint8x16_t data, uint8x16_t key) { in test_aese() 12 uint8x16_t test_aesd(uint8x16_t data, uint8x16_t key) { in test_aesd() 18 uint8x16_t test_aesmc(uint8x16_t data, uint8x16_t key) { in test_aesmc() 24 uint8x16_t test_aesimc(uint8x16_t data, uint8x16_t key) { in test_aesimc()
|
D | neon-crypto.c | 13 uint8x16_t test_vaeseq_u8(uint8x16_t data, uint8x16_t key) { in test_vaeseq_u8() 20 uint8x16_t test_vaesdq_u8(uint8x16_t data, uint8x16_t key) { in test_vaesdq_u8() 26 uint8x16_t test_vaesmcq_u8(uint8x16_t data) { in test_vaesmcq_u8() 32 uint8x16_t test_vaesimcq_u8(uint8x16_t data) { in test_vaesimcq_u8()
|
D | aarch64-neon-3v.c | 62 uint8x16_t test_vandq_u8(uint8x16_t a, uint8x16_t b) { in test_vandq_u8() 158 uint8x16_t test_vorrq_u8(uint8x16_t a, uint8x16_t b) { in test_vorrq_u8() 254 uint8x16_t test_veorq_u8(uint8x16_t a, uint8x16_t b) { in test_veorq_u8() 350 uint8x16_t test_vbicq_u8(uint8x16_t a, uint8x16_t b) { in test_vbicq_u8() 446 uint8x16_t test_vornq_u8(uint8x16_t a, uint8x16_t b) { in test_vornq_u8()
|
D | arm64_neon_high_half.c | 21 uint16x8_t test_vaddw_high_u8(uint16x8_t lhs, uint8x16_t rhs) { in test_vaddw_high_u8() 51 uint16x8_t test_vsubw_high_u8(uint16x8_t lhs, uint8x16_t rhs) { in test_vsubw_high_u8() 81 uint16x8_t test_vabdl_high_u8(uint8x16_t lhs, uint8x16_t rhs) { in test_vabdl_high_u8() 111 uint16x8_t test_vabal_high_u8(uint16x8_t accum, uint8x16_t lhs, uint8x16_t rhs) { in test_vabal_high_u8() 231 uint16x8_t test_vsubl_high_u8(uint8x16_t lhs, uint8x16_t rhs) { in test_vsubl_high_u8() 261 uint8x16_t test_vrshrn_high_n_u16(uint8x8_t lowpart, uint16x8_t input) { in test_vrshrn_high_n_u16() 291 uint8x16_t test_vshrn_high_n_u16(uint8x8_t lowpart, uint16x8_t input) { in test_vshrn_high_n_u16() 306 uint8x16_t test_vqshrun_high_n_s16(uint8x8_t lowpart, int16x8_t input) { in test_vqshrun_high_n_s16() 321 uint8x16_t test_vqrshrun_high_n_s16(uint8x8_t lowpart, int16x8_t input) { in test_vqrshrun_high_n_s16() 351 uint8x16_t test_vqshrn_high_n_u16(uint8x8_t lowpart, uint16x8_t input) { in test_vqshrn_high_n_u16() [all …]
|
D | aarch64-neon-perm.c | 56 uint8x16_t test_vuzp1q_u8(uint8x16_t a, uint8x16_t b) { in test_vuzp1q_u8() 182 uint8x16_t test_vuzp2q_u8(uint8x16_t a, uint8x16_t b) { in test_vuzp2q_u8() 308 uint8x16_t test_vzip1q_u8(uint8x16_t a, uint8x16_t b) { in test_vzip1q_u8() 434 uint8x16_t test_vzip2q_u8(uint8x16_t a, uint8x16_t b) { in test_vzip2q_u8() 560 uint8x16_t test_vtrn1q_u8(uint8x16_t a, uint8x16_t b) { in test_vtrn1q_u8() 686 uint8x16_t test_vtrn2q_u8(uint8x16_t a, uint8x16_t b) { in test_vtrn2q_u8() 837 uint8x16x2_t test_vuzpq_u8(uint8x16_t a, uint8x16_t b) { in test_vuzpq_u8() 947 uint8x16x2_t test_vzipq_u8(uint8x16_t a, uint8x16_t b) { in test_vzipq_u8() 1057 uint8x16x2_t test_vtrnq_u8(uint8x16_t a, uint8x16_t b) { in test_vtrnq_u8()
|
D | arm_neon_intrinsics.c | 68 uint8x16_t test_vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) { in test_vabaq_u8() 184 uint8x16_t test_vabdq_u8(uint8x16_t a, uint8x16_t b) { in test_vabdq_u8() 379 uint8x16_t test_vaddq_u8(uint8x16_t a, uint8x16_t b) { in test_vaddq_u8() 587 uint8x16_t test_vandq_u8(uint8x16_t a, uint8x16_t b) { in test_vandq_u8() 684 uint8x16_t test_vbicq_u8(uint8x16_t a, uint8x16_t b) { in test_vbicq_u8() 775 int8x16_t test_vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c) { in test_vbslq_s8() 799 uint8x16_t test_vbslq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) { in test_vbslq_u8() 829 poly8x16_t test_vbslq_p8(uint8x16_t a, poly8x16_t b, poly8x16_t c) { in test_vbslq_p8() 942 uint8x16_t test_vceqq_s8(int8x16_t a, int8x16_t b) { in test_vceqq_s8() 966 uint8x16_t test_vceqq_u8(uint8x16_t a, uint8x16_t b) { in test_vceqq_u8() [all …]
|
D | arm64_vshift.c | 77 uint8x16_t test_vqshlq_n_u8(uint8x16_t in) { in test_vqshlq_n_u8() 174 uint8x16_t test_vrshrq_n_u8(uint8x16_t in) { in test_vrshrq_n_u8() 331 uint8x16_t test_vrsraq_n_u8(uint8x16_t acc, uint8x16_t in) { in test_vrsraq_n_u8()
|
D | aarch64-neon-misc.c | 47 uint8x16_t test_vceqzq_s8(int8x16_t a) { in test_vceqzq_s8() 89 uint8x16_t test_vceqzq_u8(uint8x16_t a) { in test_vceqzq_u8() 137 uint8x16_t test_vceqzq_p8(poly8x16_t a) { in test_vceqzq_p8() 191 uint8x16_t test_vcgezq_s8(int8x16_t a) { in test_vcgezq_s8() 263 uint8x16_t test_vclezq_s8(int8x16_t a) { in test_vclezq_s8() 335 uint8x16_t test_vcgtzq_s8(int8x16_t a) { in test_vcgtzq_s8() 407 uint8x16_t test_vcltzq_s8(int8x16_t a) { in test_vcltzq_s8() 479 uint8x16_t test_vrev16q_u8(uint8x16_t a) { in test_vrev16q_u8() 539 uint8x16_t test_vrev32q_u8(uint8x16_t a) { in test_vrev32q_u8() 635 uint8x16_t test_vrev64q_u8(uint8x16_t a) { in test_vrev64q_u8() [all …]
|
D | aarch64-neon-intrinsics.c | 99 uint8x16_t test_vaddq_u8(uint8x16_t v1, uint8x16_t v2) { in test_vaddq_u8() 211 uint8x16_t test_vsubq_u8(uint8x16_t v1, uint8x16_t v2) { in test_vsubq_u8() 296 uint8x16_t test_vmulq_u8(uint8x16_t v1, uint8x16_t v2) { in test_vmulq_u8() 405 uint8x16_t test_vmlaq_u8(uint8x16_t v1, uint8x16_t v2, uint8x16_t v3) { in test_vmlaq_u8() 493 uint8x16_t test_vmlsq_u8(uint8x16_t v1, uint8x16_t v2, uint8x16_t v3) { in test_vmlsq_u8() 623 uint8x16_t test_vabaq_u8(uint8x16_t v1, uint8x16_t v2, uint8x16_t v3) { in test_vabaq_u8() 701 uint8x16_t test_vabdq_u8(uint8x16_t v1, uint8x16_t v2) { in test_vabdq_u8() 804 int8x16_t test_vbslq_s8(uint8x16_t v1, int8x16_t v2, int8x16_t v3) { in test_vbslq_s8() 828 uint8x16_t test_vbslq_u8(uint8x16_t v1, uint8x16_t v2, uint8x16_t v3) { in test_vbslq_u8() 858 poly8x16_t test_vbslq_p8(uint8x16_t v1, poly8x16_t v2, poly8x16_t v3) { in test_vbslq_p8() [all …]
|
D | arm64_vcopy.c | 13 uint8x16_t test_vcopyq_laneq_u8(uint8x16_t a1, uint8x16_t a2) { in test_vcopyq_laneq_u8()
|
D | aarch64-neon-across.c | 51 uint16_t test_vaddlvq_u8(uint8x16_t a) { in test_vaddlvq_u8() 111 uint8_t test_vmaxvq_u8(uint8x16_t a) { in test_vmaxvq_u8() 171 uint8_t test_vminvq_u8(uint8x16_t a) { in test_vminvq_u8() 231 uint8_t test_vaddvq_u8(uint8x16_t a) { in test_vaddvq_u8()
|
D | arm64_vsri.c | 108 uint8x16_t test_vsriq_n_u8(uint8x16_t a1, uint8x16_t a2) { in test_vsriq_n_u8()
|
/external/skia/src/opts/ |
D | SkTextureCompression_opts_neon.cpp | 19 static inline uint8x16_t convert_indices(const uint8x16_t &x) { in convert_indices() 109 static inline uint8x16_t make_index_row(const uint8x16_t &x) { in make_index_row() 110 static const uint8x16_t kTopThreeMask = { in make_index_row() 118 static inline bool is_zero(uint8x16_t x) { in is_zero() 160 const uint8x16_t alphaRow1 = vld1q_u8(src1); in compress_r11eac_blocks() 161 const uint8x16_t alphaRow2 = vld1q_u8(src2); in compress_r11eac_blocks() 162 const uint8x16_t alphaRow3 = vld1q_u8(src3); in compress_r11eac_blocks() 163 const uint8x16_t alphaRow4 = vld1q_u8(src4); in compress_r11eac_blocks() 165 const uint8x16_t cmp12 = vceqq_u8(alphaRow1, alphaRow2); in compress_r11eac_blocks() 166 const uint8x16_t cmp34 = vceqq_u8(alphaRow3, alphaRow4); in compress_r11eac_blocks() [all …]
|
D | Sk4px_NEON.h | 8 inline Sk4px::Sk4px(SkPMColor px) : INHERITED((uint8x16_t)vdupq_n_u32(px)) {} in Sk4px() 11 return Sk16b((uint8x16_t)vld1q_u32(px)); in Load4() 15 return Sk16b((uint8x16_t)vcombine_u32(px2, px2)); in Load2() 18 return Sk16b((uint8x16_t)vdupq_n_u32(*px)); in Load1() 57 return Sk16b((uint8x16_t)as); in alphas() 61 uint8x16_t a8 = vdupq_n_u8(0); // ____ ____ ____ ____ in Load4Alphas() 69 return Sk16b((uint8x16_t)a32); in Load4Alphas() 73 uint8x16_t a8 = vdupq_n_u8(0); // ____ ____ ____ ____ in Load2Alphas() 79 return Sk16b((uint8x16_t)a32); in Load2Alphas() 83 return Sk16b(vandq_u8(this->fVec, (uint8x16_t)vdupq_n_u32(0xFF << SK_A32_SHIFT))); in zeroColors() [all …]
|
D | SkBitmapProcState_arm_neon.cpp | 145 uint8x16_t pixels = vld1q_u8(rowToFilter); in convolveHorizontally_neon() 185 uint8x16_t pixels = vld1q_u8(rowToFilter); in convolveHorizontally_neon() 241 uint8x16_t src8 = vld1q_u8(&sourceDataRows[filterY][outX << 2]); in convolveVertically_neon() 270 uint8x16_t accum8 = vcombine_u8(vqmovun_s16(accum16_0), vqmovun_s16(accum16_1)); in convolveVertically_neon() 275 uint8x16_t a = vreinterpretq_u8_u32(vshrq_n_u32(vreinterpretq_u32_u8(accum8), 8)); in convolveVertically_neon() 277 uint8x16_t b = vmaxq_u8(a, accum8); // Max of r and g in convolveVertically_neon() 309 uint8x16_t src8 = vld1q_u8(&sourceDataRows[filterY][width << 2]); in convolveVertically_neon() 329 uint8x16_t accum8 = vcombine_u8(vqmovun_s16(accum16_0), vqmovun_s16(accum16_1)); in convolveVertically_neon() 334 uint8x16_t a = vreinterpretq_u8_u32(vshrq_n_u32(vreinterpretq_u32_u8(accum8), 8)); in convolveVertically_neon() 336 uint8x16_t b = vmaxq_u8(a, accum8); // Max of r and g in convolveVertically_neon() [all …]
|
/external/libvpx/libvpx/vp9/encoder/arm/neon/ |
D | vp9_sad_neon.c | 43 const uint8x16_t vec_src_00 = vld1q_u8(src); in vp9_sad64x64_neon() 44 const uint8x16_t vec_src_16 = vld1q_u8(src + 16); in vp9_sad64x64_neon() 45 const uint8x16_t vec_src_32 = vld1q_u8(src + 32); in vp9_sad64x64_neon() 46 const uint8x16_t vec_src_48 = vld1q_u8(src + 48); in vp9_sad64x64_neon() 47 const uint8x16_t vec_ref_00 = vld1q_u8(ref); in vp9_sad64x64_neon() 48 const uint8x16_t vec_ref_16 = vld1q_u8(ref + 16); in vp9_sad64x64_neon() 49 const uint8x16_t vec_ref_32 = vld1q_u8(ref + 32); in vp9_sad64x64_neon() 50 const uint8x16_t vec_ref_48 = vld1q_u8(ref + 48); in vp9_sad64x64_neon() 80 const uint8x16_t vec_src_00 = vld1q_u8(src); in vp9_sad32x32_neon() 81 const uint8x16_t vec_src_16 = vld1q_u8(src + 16); in vp9_sad32x32_neon() [all …]
|
D | vp9_subtract_neon.c | 26 const uint8x16_t v_src_00 = vld1q_u8(&src[c + 0]); in vp9_subtract_block_neon() 27 const uint8x16_t v_src_16 = vld1q_u8(&src[c + 16]); in vp9_subtract_block_neon() 28 const uint8x16_t v_pred_00 = vld1q_u8(&pred[c + 0]); in vp9_subtract_block_neon() 29 const uint8x16_t v_pred_16 = vld1q_u8(&pred[c + 16]); in vp9_subtract_block_neon() 49 const uint8x16_t v_src = vld1q_u8(&src[0]); in vp9_subtract_block_neon() 50 const uint8x16_t v_pred = vld1q_u8(&pred[0]); in vp9_subtract_block_neon()
|
/external/scrypt/lib/crypto/ |
D | crypto_scrypt-neon.c | 54 static void blockmix_salsa8(uint8x16_t *, uint8x16_t *, uint8x16_t *, size_t); 61 uint8x16_t * D = dest; in blkcpy() 62 uint8x16_t * S = src; in blkcpy() 73 uint8x16_t * D = dest; in blkxor() 74 uint8x16_t * S = src; in blkxor() 88 blockmix_salsa8(uint8x16_t * Bin, uint8x16_t * Bout, uint8x16_t * X, size_t r) in blockmix_salsa8() 136 uint8x16_t * X = XY; in smix() 137 uint8x16_t * Y = (void *)((uintptr_t)(XY) + 128 * r); in smix() 138 uint8x16_t * Z = (void *)((uintptr_t)(XY) + 256 * r); in smix()
|
D | crypto_scrypt-neon-salsa208.h | 116 vst1q_u8((uint8_t *) input,(uint8x16_t) x0x1x2x3); in salsa20_8_intrinsic() 117 vst1q_u8(16 + (uint8_t *) input,(uint8x16_t) x4x5x6x7); in salsa20_8_intrinsic() 118 vst1q_u8(32 + (uint8_t *) input,(uint8x16_t) x8x9x10x11); in salsa20_8_intrinsic() 119 vst1q_u8(48 + (uint8_t *) input,(uint8x16_t) x12x13x14x15); in salsa20_8_intrinsic()
|
/external/libvpx/libvpx/vp8/common/arm/neon/ |
D | mbloopfilter_neon.c | 15 uint8x16_t qblimit, // mblimit in vp8_mbloop_filter_neon() 16 uint8x16_t qlimit, // limit in vp8_mbloop_filter_neon() 17 uint8x16_t qthresh, // thresh in vp8_mbloop_filter_neon() 18 uint8x16_t q3, // p2 in vp8_mbloop_filter_neon() 19 uint8x16_t q4, // p2 in vp8_mbloop_filter_neon() 20 uint8x16_t q5, // p1 in vp8_mbloop_filter_neon() 21 uint8x16_t q6, // p0 in vp8_mbloop_filter_neon() 22 uint8x16_t q7, // q0 in vp8_mbloop_filter_neon() 23 uint8x16_t q8, // q1 in vp8_mbloop_filter_neon() 24 uint8x16_t q9, // q2 in vp8_mbloop_filter_neon() [all …]
|
/external/scrypt/patches/ |
D | arm-neon.patch | 122 + vst1q_u8((uint8_t *) input,(uint8x16_t) x0x1x2x3); 123 + vst1q_u8(16 + (uint8_t *) input,(uint8x16_t) x4x5x6x7); 124 + vst1q_u8(32 + (uint8_t *) input,(uint8x16_t) x8x9x10x11); 125 + vst1q_u8(48 + (uint8_t *) input,(uint8x16_t) x12x13x14x15); 186 +static void blockmix_salsa8(uint8x16_t *, uint8x16_t *, uint8x16_t *, size_t); 193 + uint8x16_t * D = dest; 194 + uint8x16_t * S = src; 205 + uint8x16_t * D = dest; 206 + uint8x16_t * S = src; 220 +blockmix_salsa8(uint8x16_t * Bin, uint8x16_t * Bout, uint8x16_t * X, size_t r) [all …]
|