/external/libjpeg-turbo/simd/arm/ |
D | jidctint-neon.c | 97 static INLINE void jsimd_idct_islow_pass1_regular(int16x4_t row0, 98 int16x4_t row1, 99 int16x4_t row2, 100 int16x4_t row3, 101 int16x4_t row4, 102 int16x4_t row5, 103 int16x4_t row6, 104 int16x4_t row7, 105 int16x4_t quant_row0, 106 int16x4_t quant_row1, [all …]
|
D | jidctfst-neon.c | 95 const int16x4_t consts = vld1_s16(jsimd_idct_ifast_neon_consts); in jsimd_idct_ifast_neon() 113 int16x4_t dcval = vget_low_s16(row0); in jsimd_idct_ifast_neon() 118 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 119 int16x4_t quant_row2 = vld1_s16(quantptr + 2 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 120 int16x4_t quant_row3 = vld1_s16(quantptr + 3 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 121 int16x4_t quant_row4 = vld1_s16(quantptr + 4 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 122 int16x4_t quant_row5 = vld1_s16(quantptr + 5 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 123 int16x4_t quant_row6 = vld1_s16(quantptr + 6 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 124 int16x4_t quant_row7 = vld1_s16(quantptr + 7 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 127 int16x4_t tmp0 = vget_high_s16(row0); in jsimd_idct_ifast_neon() [all …]
|
D | jidctred-neon.c | 103 const int16x4_t consts = vld1_s16(jsimd_idct_2x2_neon_consts); in jsimd_idct_2x2_neon() 230 const int16x4_t consts1 = vld1_s16(jsimd_idct_4x4_neon_consts); in jsimd_idct_4x4_neon() 231 const int16x4_t consts2 = vld1_s16(jsimd_idct_4x4_neon_consts + 4); in jsimd_idct_4x4_neon() 232 const int16x4_t consts3 = vld1_s16(jsimd_idct_4x4_neon_consts + 8); in jsimd_idct_4x4_neon() 249 int16x4_t dcval = vshl_n_s16(vget_low_s16(row0), PASS1_BITS); in jsimd_idct_4x4_neon() 254 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 255 int16x4_t quant_row2 = vld1_s16(quantptr + 2 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 256 int16x4_t quant_row3 = vld1_s16(quantptr + 3 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 257 int16x4_t quant_row5 = vld1_s16(quantptr + 5 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 258 int16x4_t quant_row6 = vld1_s16(quantptr + 6 * DCTSIZE + 4); in jsimd_idct_4x4_neon() [all …]
|
/external/webrtc/common_audio/signal_processing/ |
D | downsample_fast_neon.c | 51 int16x4_t coeff16x4 = vreinterpret_s16_s32(coeff32); in WebRtcSpl_DownsampleFastNeon() 55 int16x4_t in16x4_0 = vget_low_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 56 int16x4_t in16x4_1 = vget_low_s16(in16x8x2.val[1]); in WebRtcSpl_DownsampleFastNeon() 65 int16x4_t in16x4_2 = vget_high_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 66 int16x4_t in16x4_3 = vget_high_s16(in16x8x2.val[1]); in WebRtcSpl_DownsampleFastNeon() 72 int16x4_t coeff16x4 = vld1_dup_s16(&coefficients[j]); in WebRtcSpl_DownsampleFastNeon() 76 int16x4_t in16x4_0 = vget_low_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 84 int16x4_t in16x4_1 = vget_high_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 91 int16x4_t coeff16x4 = vld1_dup_s16(&coefficients[j]); in WebRtcSpl_DownsampleFastNeon() 95 int16x4_t in16x4_0 = vget_low_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() [all …]
|
/external/clang/test/CodeGen/ |
D | arm-v8.1a-neon-intrinsics.c | 14 int16x4_t test_vqrdmlah_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlah_s16() 54 int16x4_t test_vqrdmlah_lane_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlah_lane_s16() 78 int16x8_t test_vqrdmlahq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t c) { in test_vqrdmlahq_lane_s16() 102 int16x4_t test_vqrdmlsh_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlsh_s16() 142 int16x4_t test_vqrdmlsh_lane_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlsh_lane_s16() 166 int16x8_t test_vqrdmlshq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t c) { in test_vqrdmlshq_lane_s16()
|
D | aarch64-neon-2velem.c | 12 int16x4_t test_vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v) { in test_vmla_lane_s16() 21 int16x8_t test_vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v) { in test_vmlaq_lane_s16() 48 int16x4_t test_vmla_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vmla_laneq_s16() 84 int16x4_t test_vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v) { in test_vmls_lane_s16() 93 int16x8_t test_vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v) { in test_vmlsq_lane_s16() 120 int16x4_t test_vmls_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vmls_laneq_s16() 155 int16x4_t test_vmul_lane_s16(int16x4_t a, int16x4_t v) { in test_vmul_lane_s16() 163 int16x8_t test_vmulq_lane_s16(int16x8_t a, int16x4_t v) { in test_vmulq_lane_s16() 219 int16x4_t test_vmul_laneq_s16(int16x4_t a, int16x8_t v) { in test_vmul_laneq_s16() 505 int32x4_t test_vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v) { in test_vmlal_lane_s16() [all …]
|
D | arm_neon_intrinsics.c | 27 int16x4_t test_vaba_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vaba_s16() 176 int32x4_t test_vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c) { in test_vabal_s16() 257 int16x4_t test_vabd_s16(int16x4_t a, int16x4_t b) { in test_vabd_s16() 420 int32x4_t test_vabdl_s16(int16x4_t a, int16x4_t b) { in test_vabdl_s16() 493 int16x4_t test_vabs_s16(int16x4_t a) { in test_vabs_s16() 560 int16x4_t test_vadd_s16(int16x4_t a, int16x4_t b) { in test_vadd_s16() 699 int16x4_t test_vaddhn_s32(int32x4_t a, int32x4_t b) { in test_vaddhn_s32() 774 int32x4_t test_vaddl_s16(int16x4_t a, int16x4_t b) { in test_vaddl_s16() 841 int32x4_t test_vaddw_s16(int32x4_t a, int16x4_t b) { in test_vaddw_s16() 894 int16x4_t test_vand_s16(int16x4_t a, int16x4_t b) { in test_vand_s16() [all …]
|
D | aarch64-v8.1a-neon-intrinsics.c | 9 int16x4_t test_vqrdmlah_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vqrdmlah_laneq_s16() 61 int16_t test_vqrdmlahh_lane_s16(int16_t a, int16_t b, int16x4_t c) { in test_vqrdmlahh_lane_s16() 105 int16x4_t test_vqrdmlsh_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vqrdmlsh_laneq_s16() 157 int16_t test_vqrdmlshh_lane_s16(int16_t a, int16_t b, int16x4_t c) { in test_vqrdmlshh_lane_s16()
|
/external/llvm-project/clang/test/CodeGen/ |
D | arm-v8.1a-neon-intrinsics.c | 14 int16x4_t test_vqrdmlah_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlah_s16() 54 int16x4_t test_vqrdmlah_lane_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlah_lane_s16() 78 int16x8_t test_vqrdmlahq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t c) { in test_vqrdmlahq_lane_s16() 102 int16x4_t test_vqrdmlsh_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlsh_s16() 142 int16x4_t test_vqrdmlsh_lane_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vqrdmlsh_lane_s16() 166 int16x8_t test_vqrdmlshq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t c) { in test_vqrdmlshq_lane_s16()
|
D | aarch64-neon-2velem.c | 17 int16x4_t test_vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v) { in test_vmla_lane_s16() 30 int16x8_t test_vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v) { in test_vmlaq_lane_s16() 69 int16x4_t test_vmla_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vmla_laneq_s16() 121 int16x4_t test_vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v) { in test_vmls_lane_s16() 134 int16x8_t test_vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v) { in test_vmlsq_lane_s16() 173 int16x4_t test_vmls_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vmls_laneq_s16() 224 int16x4_t test_vmul_lane_s16(int16x4_t a, int16x4_t v) { in test_vmul_lane_s16() 236 int16x8_t test_vmulq_lane_s16(int16x8_t a, int16x4_t v) { in test_vmulq_lane_s16() 320 int16x4_t test_vmul_laneq_s16(int16x4_t a, int16x8_t v) { in test_vmul_laneq_s16() 660 int32x4_t test_vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v) { in test_vmlal_lane_s16() [all …]
|
D | arm_neon_intrinsics.c | 25 int16x4_t test_vaba_s16(int16x4_t a, int16x4_t b, int16x4_t c) { in test_vaba_s16() 148 int32x4_t test_vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c) { in test_vabal_s16() 213 int16x4_t test_vabd_s16(int16x4_t a, int16x4_t b) { in test_vabd_s16() 344 int32x4_t test_vabdl_s16(int16x4_t a, int16x4_t b) { in test_vabdl_s16() 403 int16x4_t test_vabs_s16(int16x4_t a) { in test_vabs_s16() 464 int16x4_t test_vadd_s16(int16x4_t a, int16x4_t b) { in test_vadd_s16() 598 int16x4_t test_vaddhn_s32(int32x4_t a, int32x4_t b) { in test_vaddhn_s32() 662 int32x4_t test_vaddl_s16(int16x4_t a, int16x4_t b) { in test_vaddl_s16() 721 int32x4_t test_vaddw_s16(int32x4_t a, int16x4_t b) { in test_vaddw_s16() 770 int16x4_t test_vand_s16(int16x4_t a, int16x4_t b) { in test_vand_s16() [all …]
|
D | aarch64-neon-misc.c | 22 uint16x4_t test_vceqz_s16(int16x4_t a) { in test_vceqz_s16() 250 uint16x4_t test_vcgez_s16(int16x4_t a) { in test_vcgez_s16() 356 uint16x4_t test_vclez_s16(int16x4_t a) { in test_vclez_s16() 462 uint16x4_t test_vcgtz_s16(int16x4_t a) { in test_vcgtz_s16() 568 uint16x4_t test_vcltz_s16(int16x4_t a) { in test_vcltz_s16() 713 int16x4_t test_vrev32_s16(int16x4_t a) { in test_vrev32_s16() 797 int16x4_t test_vrev64_s16(int16x4_t a) { in test_vrev64_s16() 916 int16x4_t test_vpaddl_s8(int8x8_t a) { in test_vpaddl_s8() 924 int32x2_t test_vpaddl_s16(int16x4_t a) { in test_vpaddl_s16() 1010 int16x4_t test_vpadal_s8(int16x4_t a, int8x8_t b) { in test_vpadal_s8() [all …]
|
D | aarch64-v8.1a-neon-intrinsics.c | 9 int16x4_t test_vqrdmlah_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vqrdmlah_laneq_s16() 61 int16_t test_vqrdmlahh_lane_s16(int16_t a, int16_t b, int16x4_t c) { in test_vqrdmlahh_lane_s16() 105 int16x4_t test_vqrdmlsh_laneq_s16(int16x4_t a, int16x4_t b, int16x8_t v) { in test_vqrdmlsh_laneq_s16() 157 int16_t test_vqrdmlshh_lane_s16(int16_t a, int16_t b, int16x4_t c) { in test_vqrdmlshh_lane_s16()
|
/external/libaom/libaom/av1/common/arm/ |
D | convolve_neon.h | 118 const int16x4_t s0, const int16x4_t s1, const int16x4_t s2, in wiener_convolve8_horiz_4x8() 119 const int16x4_t s3, const int16x4_t s4, const int16x4_t s5, in wiener_convolve8_horiz_4x8() 120 const int16x4_t s6, int16_t *filter_x, const int bd, in wiener_convolve8_horiz_4x8() 124 int16x4_t sum, temp0, temp1, temp2; in wiener_convolve8_horiz_4x8() 181 static INLINE int16x4_t 182 convolve8_4x4_s16(const int16x4_t s0, const int16x4_t s1, const int16x4_t s2, in convolve8_4x4_s16() 183 const int16x4_t s3, const int16x4_t s4, const int16x4_t s5, in convolve8_4x4_s16() 184 const int16x4_t s6, const int16x4_t s7, const int16_t *filter, in convolve8_4x4_s16() 185 const int16x4_t horiz_const, const int16x4_t shift_round_0) { in convolve8_4x4_s16() 186 int16x4_t sum; in convolve8_4x4_s16() [all …]
|
/external/libvpx/libvpx/vpx_dsp/arm/ |
D | fdct_neon.c | 25 int16x4_t input_0 = vshl_n_s16(vld1_s16(input + 0 * stride), 4); in vpx_fdct4x4_neon() 26 int16x4_t input_1 = vshl_n_s16(vld1_s16(input + 1 * stride), 4); in vpx_fdct4x4_neon() 27 int16x4_t input_2 = vshl_n_s16(vld1_s16(input + 2 * stride), 4); in vpx_fdct4x4_neon() 28 int16x4_t input_3 = vshl_n_s16(vld1_s16(input + 3 * stride), 4); in vpx_fdct4x4_neon() 32 const int16x4_t one = vreinterpret_s16_s64(vdup_n_s64(1)); in vpx_fdct4x4_neon() 45 const int16x4_t s_0 = vget_low_s16(s_01); in vpx_fdct4x4_neon() 46 const int16x4_t s_1 = vget_high_s16(s_01); in vpx_fdct4x4_neon() 47 const int16x4_t s_2 = vget_high_s16(s_32); in vpx_fdct4x4_neon() 48 const int16x4_t s_3 = vget_low_s16(s_32); in vpx_fdct4x4_neon() 58 int16x4_t out_0 = vrshrn_n_s32(temp1, DCT_CONST_BITS); in vpx_fdct4x4_neon() [all …]
|
D | vpx_convolve8_neon.h | 75 static INLINE int16x4_t convolve8_4(const int16x4_t s0, const int16x4_t s1, in convolve8_4() 76 const int16x4_t s2, const int16x4_t s3, in convolve8_4() 77 const int16x4_t s4, const int16x4_t s5, in convolve8_4() 78 const int16x4_t s6, const int16x4_t s7, in convolve8_4() 80 const int16x4_t filter3, in convolve8_4() 81 const int16x4_t filter4) { in convolve8_4() 82 const int16x4_t filters_lo = vget_low_s16(filters); in convolve8_4() 83 const int16x4_t filters_hi = vget_high_s16(filters); in convolve8_4() 84 int16x4_t sum; in convolve8_4() 104 const int16x4_t filters_lo = vget_low_s16(filters); in convolve8_8() [all …]
|
D | fwd_txfm_neon.c | 65 const int16x4_t a = vrshrn_n_s32(v_t0_lo, DCT_CONST_BITS); in vpx_fdct8x8_neon() 66 const int16x4_t b = vrshrn_n_s32(v_t0_hi, DCT_CONST_BITS); in vpx_fdct8x8_neon() 67 const int16x4_t c = vrshrn_n_s32(v_t1_lo, DCT_CONST_BITS); in vpx_fdct8x8_neon() 68 const int16x4_t d = vrshrn_n_s32(v_t1_hi, DCT_CONST_BITS); in vpx_fdct8x8_neon() 69 const int16x4_t e = vrshrn_n_s32(v_t2_lo, DCT_CONST_BITS); in vpx_fdct8x8_neon() 70 const int16x4_t f = vrshrn_n_s32(v_t2_hi, DCT_CONST_BITS); in vpx_fdct8x8_neon() 71 const int16x4_t g = vrshrn_n_s32(v_t3_lo, DCT_CONST_BITS); in vpx_fdct8x8_neon() 72 const int16x4_t h = vrshrn_n_s32(v_t3_hi, DCT_CONST_BITS); in vpx_fdct8x8_neon() 86 const int16x4_t a = vrshrn_n_s32(v_t0_lo, DCT_CONST_BITS); in vpx_fdct8x8_neon() 87 const int16x4_t b = vrshrn_n_s32(v_t0_hi, DCT_CONST_BITS); in vpx_fdct8x8_neon() [all …]
|
D | idct16x16_add_neon.c | 18 static INLINE void wrap_low_4x2(const int32x4_t *const t32, int16x4_t *const d0, in wrap_low_4x2() 19 int16x4_t *const d1) { in wrap_low_4x2() 24 static INLINE void idct_cospi_8_24_d_kernel(const int16x4_t s0, in idct_cospi_8_24_d_kernel() 25 const int16x4_t s1, in idct_cospi_8_24_d_kernel() 26 const int16x4_t cospi_0_8_16_24, in idct_cospi_8_24_d_kernel() 34 static INLINE void idct_cospi_8_24_d(const int16x4_t s0, const int16x4_t s1, in idct_cospi_8_24_d() 35 const int16x4_t cospi_0_8_16_24, in idct_cospi_8_24_d() 36 int16x4_t *const d0, int16x4_t *const d1) { in idct_cospi_8_24_d() 43 static INLINE void idct_cospi_8_24_neg_d(const int16x4_t s0, const int16x4_t s1, in idct_cospi_8_24_neg_d() 44 const int16x4_t cospi_0_8_16_24, in idct_cospi_8_24_neg_d() [all …]
|
D | idct8x8_add_neon.c | 23 const int16x4_t cospis0 = vget_low_s16(cospis); // cospi 0, 8, 16, 24 in vpx_idct8x8_64_add_neon() 24 const int16x4_t cospis1 = vget_high_s16(cospis); // cospi 4, 12, 20, 28 in vpx_idct8x8_64_add_neon() 45 const int16x4_t cospis0 = vget_low_s16(cospis); // cospi 0, 8, 16, 24 in vpx_idct8x8_12_add_neon() 46 const int16x4_t cospisd0 = vget_low_s16(cospisd); // doubled 0, 8, 16, 24 in vpx_idct8x8_12_add_neon() 47 const int16x4_t cospisd1 = vget_high_s16(cospisd); // doubled 4, 12, 20, 28 in vpx_idct8x8_12_add_neon() 48 int16x4_t a[8]; in vpx_idct8x8_12_add_neon()
|
/external/libaom/libaom/aom_dsp/arm/ |
D | fwd_txfm_neon.c | 21 int16x4_t *input_0, int16x4_t *input_1, in aom_fdct4x4_helper() 22 int16x4_t *input_2, int16x4_t *input_3) { in aom_fdct4x4_helper() 29 const int16x4_t one = vreinterpret_s16_s64(vdup_n_s64(1)); in aom_fdct4x4_helper() 42 const int16x4_t s_0 = vget_low_s16(s_01); in aom_fdct4x4_helper() 43 const int16x4_t s_1 = vget_high_s16(s_01); in aom_fdct4x4_helper() 44 const int16x4_t s_2 = vget_high_s16(s_32); in aom_fdct4x4_helper() 45 const int16x4_t s_3 = vget_low_s16(s_32); in aom_fdct4x4_helper() 55 int16x4_t out_0 = vrshrn_n_s32(temp1, DCT_CONST_BITS); in aom_fdct4x4_helper() 56 int16x4_t out_2 = vrshrn_n_s32(temp2, DCT_CONST_BITS); in aom_fdct4x4_helper() 67 int16x4_t out_1 = vrshrn_n_s32(temp3, DCT_CONST_BITS); in aom_fdct4x4_helper() [all …]
|
/external/libgav1/libgav1/src/dsp/arm/ |
D | motion_vector_search_neon.cc | 36 inline int16x4_t MvProjection(const int16x4_t mv, const int16x4_t denominator, in MvProjection() 45 inline int16x4_t MvProjectionCompound(const int16x4_t mv, in MvProjectionCompound() 48 const int16x4_t denominator = in MvProjectionCompound() 56 inline int16x8_t ProjectionClip(const int16x4_t mv0, const int16x4_t mv1) { in ProjectionClip() 69 const int16x4_t tmv0 = vreinterpret_s16_s32(vdup_lane_s32(temporal_mv, 0)); in MvProjectionCompoundClip() 70 const int16x4_t tmv1 = vreinterpret_s16_s32(vdup_lane_s32(temporal_mv, 1)); in MvProjectionCompoundClip() 71 const int16x4_t mv0 = MvProjectionCompound( in MvProjectionCompoundClip() 73 const int16x4_t mv1 = MvProjectionCompound( in MvProjectionCompoundClip() 81 int16x4_t* const lookup) { in MvProjectionSingleClip() 93 const int16x4_t tmv0 = vget_low_s16(temporal_mv); in MvProjectionSingleClip() [all …]
|
/external/libopus/celt/arm/ |
D | celt_neon_intr.c | 47 int16x4_t y0 = vld1_s16(y); in xcorr_kernel_neon_fixed() 54 int16x4_t x0 = vget_low_s16(xx); in xcorr_kernel_neon_fixed() 55 int16x4_t x4 = vget_high_s16(xx); in xcorr_kernel_neon_fixed() 58 int16x4_t y4 = vget_low_s16(yy); in xcorr_kernel_neon_fixed() 59 int16x4_t y8 = vget_high_s16(yy); in xcorr_kernel_neon_fixed() 63 int16x4_t y1 = vext_s16(y0, y4, 1); in xcorr_kernel_neon_fixed() 64 int16x4_t y5 = vext_s16(y4, y8, 1); in xcorr_kernel_neon_fixed() 68 int16x4_t y2 = vext_s16(y0, y4, 2); in xcorr_kernel_neon_fixed() 69 int16x4_t y6 = vext_s16(y4, y8, 2); in xcorr_kernel_neon_fixed() 73 int16x4_t y3 = vext_s16(y0, y4, 3); in xcorr_kernel_neon_fixed() [all …]
|
/external/tensorflow/tensorflow/core/kernels/ |
D | quantized_mul_op.cc | 57 const int16x4_t scalar_minus_offset_16x4 = vmov_n_s16(scalar_minus_offset); in ScalarMultiply() 78 const int16x4_t x_high_high_16x4 = in ScalarMultiply() 80 const int16x4_t x_high_low_16x4 = in ScalarMultiply() 82 const int16x4_t x_low_high_16x4 = in ScalarMultiply() 84 const int16x4_t x_low_low_16x4 = in ScalarMultiply() 156 const int16x4_t x_high_high_16x4 = vget_high_s16(x_minus_offset_high_16x8); in VectorMultiply() 157 const int16x4_t x_high_low_16x4 = vget_low_s16(x_minus_offset_high_16x8); in VectorMultiply() 158 const int16x4_t x_low_high_16x4 = vget_high_s16(x_minus_offset_low_16x8); in VectorMultiply() 159 const int16x4_t x_low_low_16x4 = vget_low_s16(x_minus_offset_low_16x8); in VectorMultiply() 160 const int16x4_t y_high_high_16x4 = vget_high_s16(y_minus_offset_high_16x8); in VectorMultiply() [all …]
|
/external/libhevc/common/arm/ |
D | ihevc_weighted_pred_neon_intr.c | 116 int16x4_t pi2_src_val1; in ihevc_weighted_pred_uni_neonintr() 117 int16x4_t pi2_src_val2; in ihevc_weighted_pred_uni_neonintr() 249 int16x4_t pi2_src_val1; in ihevc_weighted_pred_chroma_uni_neonintr() 250 int16x4_t pi2_src_val2; in ihevc_weighted_pred_chroma_uni_neonintr() 261 int16x4_t tmp_wgt0_u, tmp_wgt0_v; in ihevc_weighted_pred_chroma_uni_neonintr() 412 int16x4_t pi2_src1_val1; in ihevc_weighted_pred_bi_neonintr() 413 int16x4_t pi2_src1_val2; in ihevc_weighted_pred_bi_neonintr() 414 int16x4_t pi2_src2_val1; in ihevc_weighted_pred_bi_neonintr() 415 int16x4_t pi2_src2_val2; in ihevc_weighted_pred_bi_neonintr() 581 int16x4_t pi2_src1_val1; in ihevc_weighted_pred_chroma_bi_neonintr() [all …]
|
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/integer_ops/ |
D | add.h | 80 const int16x4_t input1_val_high_high = vget_high_s16(input1_val_high); in AddElementwise() 81 const int16x4_t input1_val_high_low = vget_low_s16(input1_val_high); in AddElementwise() 82 const int16x4_t input1_val_low_high = vget_high_s16(input1_val_low); in AddElementwise() 83 const int16x4_t input1_val_low_low = vget_low_s16(input1_val_low); in AddElementwise() 84 const int16x4_t input2_val_high_high = vget_high_s16(input2_val_high); in AddElementwise() 85 const int16x4_t input2_val_high_low = vget_low_s16(input2_val_high); in AddElementwise() 86 const int16x4_t input2_val_low_high = vget_high_s16(input2_val_low); in AddElementwise() 87 const int16x4_t input2_val_low_low = vget_low_s16(input2_val_low); in AddElementwise() 126 const int16x4_t s11_narrowed = vmovn_s32(s11); in AddElementwise() 127 const int16x4_t s12_narrowed = vmovn_s32(s12); in AddElementwise() [all …]
|