| /external/libopus/silk/arm/ |
| D | NSQ_neon.c | 40 int32x4_t coef0 = vld1q_s32(coef32); in silk_noise_shape_quantizer_short_prediction_neon()
41 int32x4_t coef1 = vld1q_s32(coef32 + 4); in silk_noise_shape_quantizer_short_prediction_neon()
42 int32x4_t coef2 = vld1q_s32(coef32 + 8); in silk_noise_shape_quantizer_short_prediction_neon()
43 int32x4_t coef3 = vld1q_s32(coef32 + 12); in silk_noise_shape_quantizer_short_prediction_neon()
45 int32x4_t a0 = vld1q_s32(buf32 - 15); in silk_noise_shape_quantizer_short_prediction_neon()
46 int32x4_t a1 = vld1q_s32(buf32 - 11); in silk_noise_shape_quantizer_short_prediction_neon()
47 int32x4_t a2 = vld1q_s32(buf32 - 7); in silk_noise_shape_quantizer_short_prediction_neon()
48 int32x4_t a3 = vld1q_s32(buf32 - 3); in silk_noise_shape_quantizer_short_prediction_neon()
78 int32x4_t a01 = vld1q_s32(data1); /* data1[0] ... [3] */ in silk_NSQ_noise_shape_feedback_loop_neon()
81 int32x4_t a1 = vld1q_s32(data1 + 3); /* data1[3] ... [6] */ in silk_NSQ_noise_shape_feedback_loop_neon()
|
| D | NSQ_del_dec_neon_intr.c | 362 RD_Q10_s32x4 = vld1q_s32( psDelDec->RD_Q10 ); in silk_NSQ_del_dec_neon()
513 const int32x4_t a_Q12_arch0_s32x4 = vld1q_s32( a_Q12_arch + 0 ); in silk_noise_shape_quantizer_short_prediction_neon_local()
514 const int32x4_t a_Q12_arch1_s32x4 = vld1q_s32( a_Q12_arch + 4 ); in silk_noise_shape_quantizer_short_prediction_neon_local()
515 const int32x4_t a_Q12_arch2_s32x4 = vld1q_s32( a_Q12_arch + 8 ); in silk_noise_shape_quantizer_short_prediction_neon_local()
516 const int32x4_t a_Q12_arch3_s32x4 = vld1q_s32( a_Q12_arch + 12 ); in silk_noise_shape_quantizer_short_prediction_neon_local()
522 …LPC_pred_Q14_s32x4 = silk_SMLAWB_lane0_neon( LPC_pred_Q14_s32x4, vld1q_s32( buf32 + 0 * NEON_MAX_… in silk_noise_shape_quantizer_short_prediction_neon_local()
523 …LPC_pred_Q14_s32x4 = silk_SMLAWB_lane1_neon( LPC_pred_Q14_s32x4, vld1q_s32( buf32 + 1 * NEON_MAX_… in silk_noise_shape_quantizer_short_prediction_neon_local()
524 …LPC_pred_Q14_s32x4 = silk_SMLAWB_lane0_neon( LPC_pred_Q14_s32x4, vld1q_s32( buf32 + 2 * NEON_MAX_… in silk_noise_shape_quantizer_short_prediction_neon_local()
525 …LPC_pred_Q14_s32x4 = silk_SMLAWB_lane1_neon( LPC_pred_Q14_s32x4, vld1q_s32( buf32 + 3 * NEON_MAX_… in silk_noise_shape_quantizer_short_prediction_neon_local()
526 …LPC_pred_Q14_s32x4 = silk_SMLAWB_lane0_neon( LPC_pred_Q14_s32x4, vld1q_s32( buf32 + 4 * NEON_MAX_… in silk_noise_shape_quantizer_short_prediction_neon_local()
[all …]
|
| /external/libvpx/libvpx/vp9/common/arm/neon/ |
| D | vp9_highbd_iht16x16_add_neon.c | 174 in[0].val[0] = vld1q_s32(input); in highbd_iadst16_neon()
175 in[0].val[1] = vld1q_s32(input + 4); in highbd_iadst16_neon()
177 in[8].val[0] = vld1q_s32(input); in highbd_iadst16_neon()
178 in[8].val[1] = vld1q_s32(input + 4); in highbd_iadst16_neon()
180 in[1].val[0] = vld1q_s32(input); in highbd_iadst16_neon()
181 in[1].val[1] = vld1q_s32(input + 4); in highbd_iadst16_neon()
183 in[9].val[0] = vld1q_s32(input); in highbd_iadst16_neon()
184 in[9].val[1] = vld1q_s32(input + 4); in highbd_iadst16_neon()
186 in[2].val[0] = vld1q_s32(input); in highbd_iadst16_neon()
187 in[2].val[1] = vld1q_s32(input + 4); in highbd_iadst16_neon()
[all …]
|
| D | vp9_highbd_iht8x8_add_neon.c | 167 a[0] = vld1q_s32(input); in vp9_highbd_iht8x8_64_add_neon()
168 a[1] = vld1q_s32(input + 4); in vp9_highbd_iht8x8_64_add_neon()
169 a[2] = vld1q_s32(input + 8); in vp9_highbd_iht8x8_64_add_neon()
170 a[3] = vld1q_s32(input + 12); in vp9_highbd_iht8x8_64_add_neon()
171 a[4] = vld1q_s32(input + 16); in vp9_highbd_iht8x8_64_add_neon()
172 a[5] = vld1q_s32(input + 20); in vp9_highbd_iht8x8_64_add_neon()
173 a[6] = vld1q_s32(input + 24); in vp9_highbd_iht8x8_64_add_neon()
174 a[7] = vld1q_s32(input + 28); in vp9_highbd_iht8x8_64_add_neon()
175 a[8] = vld1q_s32(input + 32); in vp9_highbd_iht8x8_64_add_neon()
176 a[9] = vld1q_s32(input + 36); in vp9_highbd_iht8x8_64_add_neon()
[all …]
|
| D | vp9_highbd_iht4x4_add_neon.c | 25 const int32x4_t sinpi = vld1q_s32(sinpis); in highbd_iadst4()
83 c[0] = vld1q_s32(input); in vp9_highbd_iht4x4_16_add_neon()
84 c[1] = vld1q_s32(input + 4); in vp9_highbd_iht4x4_16_add_neon()
85 c[2] = vld1q_s32(input + 8); in vp9_highbd_iht4x4_16_add_neon()
86 c[3] = vld1q_s32(input + 12); in vp9_highbd_iht4x4_16_add_neon()
128 const int32x4_t cospis = vld1q_s32(kCospi32); in vp9_highbd_iht4x4_16_add_neon()
141 const int32x4_t cospis = vld1q_s32(kCospi32); in vp9_highbd_iht4x4_16_add_neon()
154 const int32x4_t cospis = vld1q_s32(kCospi32); in vp9_highbd_iht4x4_16_add_neon()
|
| /external/libvpx/libvpx/vpx_dsp/arm/ |
| D | highbd_idct16x16_add_neon.c | 522 const int32x4_t cospi_0_8_16_24 = vld1q_s32(kCospi32 + 0); in vpx_highbd_idct16x16_256_add_half1d()
523 const int32x4_t cospi_4_12_20N_28 = vld1q_s32(kCospi32 + 4); in vpx_highbd_idct16x16_256_add_half1d()
524 const int32x4_t cospi_2_30_10_22 = vld1q_s32(kCospi32 + 8); in vpx_highbd_idct16x16_256_add_half1d()
525 const int32x4_t cospi_6_26N_14_18N = vld1q_s32(kCospi32 + 12); in vpx_highbd_idct16x16_256_add_half1d()
529 in[0].val[0] = vld1q_s32(input); in vpx_highbd_idct16x16_256_add_half1d()
530 in[0].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct16x16_256_add_half1d()
532 in[8].val[0] = vld1q_s32(input); in vpx_highbd_idct16x16_256_add_half1d()
533 in[8].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct16x16_256_add_half1d()
535 in[1].val[0] = vld1q_s32(input); in vpx_highbd_idct16x16_256_add_half1d()
536 in[1].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct16x16_256_add_half1d()
[all …]
|
| D | highbd_idct8x8_add_neon.c | 231 a[0] = vld1q_s32(input); in vpx_highbd_idct8x8_12_add_neon()
232 a[1] = vld1q_s32(input + 8); in vpx_highbd_idct8x8_12_add_neon()
233 a[2] = vld1q_s32(input + 16); in vpx_highbd_idct8x8_12_add_neon()
234 a[3] = vld1q_s32(input + 24); in vpx_highbd_idct8x8_12_add_neon()
260 const int32x4_t cospis0 = vld1q_s32(kCospi32); // cospi 0, 8, 16, 24 in vpx_highbd_idct8x8_12_add_neon()
261 const int32x4_t cospis1 = vld1q_s32(kCospi32 + 4); // cospi 4, 12, 20, 28 in vpx_highbd_idct8x8_12_add_neon()
295 a[0] = vld1q_s32(input); in vpx_highbd_idct8x8_64_add_neon()
296 a[1] = vld1q_s32(input + 4); in vpx_highbd_idct8x8_64_add_neon()
297 a[2] = vld1q_s32(input + 8); in vpx_highbd_idct8x8_64_add_neon()
298 a[3] = vld1q_s32(input + 12); in vpx_highbd_idct8x8_64_add_neon()
[all …]
|
| D | highbd_idct4x4_add_neon.c | 57 c[0] = vld1q_s32(input); in vpx_highbd_idct4x4_16_add_neon()
58 c[1] = vld1q_s32(input + 4); in vpx_highbd_idct4x4_16_add_neon()
59 c[2] = vld1q_s32(input + 8); in vpx_highbd_idct4x4_16_add_neon()
60 c[3] = vld1q_s32(input + 12); in vpx_highbd_idct4x4_16_add_neon()
74 const int32x4_t cospis = vld1q_s32(kCospi32); in vpx_highbd_idct4x4_16_add_neon()
|
| D | highbd_idct32x32_1024_add_neon.c | 23 q0->val[0] = vld1q_s32(trans_buf + first * 8); in load_from_transformed()
24 q0->val[1] = vld1q_s32(trans_buf + first * 8 + 4); in load_from_transformed()
25 q1->val[0] = vld1q_s32(trans_buf + second * 8); in load_from_transformed()
26 q1->val[1] = vld1q_s32(trans_buf + second * 8 + 4); in load_from_transformed()
32 q0->val[0] = vld1q_s32(out + first * 32); in load_from_output()
33 q0->val[1] = vld1q_s32(out + first * 32 + 4); in load_from_output()
34 q1->val[0] = vld1q_s32(out + second * 32); in load_from_output()
35 q1->val[1] = vld1q_s32(out + second * 32 + 4); in load_from_output()
128 s[0].val[0] = vld1q_s32(in); in load_s32x4q_dual()
129 s[0].val[1] = vld1q_s32(in + 4); in load_s32x4q_dual()
[all …]
|
| D | highbd_idct32x32_135_add_neon.c | 24 in0->val[0] = vld1q_s32(input); in load_8x8_s32_dual()
25 in0->val[1] = vld1q_s32(input + 4); in load_8x8_s32_dual()
27 in1->val[0] = vld1q_s32(input); in load_8x8_s32_dual()
28 in1->val[1] = vld1q_s32(input + 4); in load_8x8_s32_dual()
30 in2->val[0] = vld1q_s32(input); in load_8x8_s32_dual()
31 in2->val[1] = vld1q_s32(input + 4); in load_8x8_s32_dual()
33 in3->val[0] = vld1q_s32(input); in load_8x8_s32_dual()
34 in3->val[1] = vld1q_s32(input + 4); in load_8x8_s32_dual()
36 in4->val[0] = vld1q_s32(input); in load_8x8_s32_dual()
37 in4->val[1] = vld1q_s32(input + 4); in load_8x8_s32_dual()
[all …]
|
| D | highbd_idct32x32_34_add_neon.c | 42 in[0].val[0] = vld1q_s32(input); in vpx_highbd_idct32_6_neon()
43 in[0].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct32_6_neon()
45 in[1].val[0] = vld1q_s32(input); in vpx_highbd_idct32_6_neon()
46 in[1].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct32_6_neon()
48 in[2].val[0] = vld1q_s32(input); in vpx_highbd_idct32_6_neon()
49 in[2].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct32_6_neon()
51 in[3].val[0] = vld1q_s32(input); in vpx_highbd_idct32_6_neon()
52 in[3].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct32_6_neon()
54 in[4].val[0] = vld1q_s32(input); in vpx_highbd_idct32_6_neon()
55 in[4].val[1] = vld1q_s32(input + 4); in vpx_highbd_idct32_6_neon()
[all …]
|
| /external/tensorflow/tensorflow/lite/kernels/internal/optimized/integer_ops/ |
| D | depthwise_conv.h | 65 acc[i].val[0] = vld1q_s32(acc_buffer_ptr + 4 * i);
66 acc[i].val[1] = vld1q_s32(acc_buffer_ptr + 4 * i + 8);
108 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
140 acc[0] = vld1q_s32(acc_buffer_ptr);
141 acc[1] = vld1q_s32(acc_buffer_ptr + 4);
175 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
202 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
245 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
276 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
321 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
[all …]
|
| /external/tensorflow/tensorflow/lite/kernels/internal/optimized/ |
| D | depthwiseconv_uint8.h | 56 acc[i].val[0] = vld1q_s32(acc_buffer_ptr + 4 * i);
57 acc[i].val[1] = vld1q_s32(acc_buffer_ptr + 4 * i + 8);
100 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
132 acc[0] = vld1q_s32(acc_buffer_ptr);
133 acc[1] = vld1q_s32(acc_buffer_ptr + 4);
168 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
195 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
240 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
272 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
320 acc[i] = vld1q_s32(acc_buffer_ptr + 4 * i);
[all …]
|
| /external/webrtc/modules/audio_coding/codecs/isac/fix/source/ |
| D | lattice_neon.c | 47 ptr0va = vld1q_s32(ptr0); in WebRtcIsacfix_FilterMaLoopNeon()
48 ptr0vb = vld1q_s32(ptr0 + 4); in WebRtcIsacfix_FilterMaLoopNeon()
51 ptr2va = vld1q_s32(ptr2); in WebRtcIsacfix_FilterMaLoopNeon()
52 ptr2vb = vld1q_s32(ptr2 + 4); in WebRtcIsacfix_FilterMaLoopNeon()
104 ptr0va = vld1q_s32(ptr0); in WebRtcIsacfix_FilterMaLoopNeon()
105 ptr2va = vld1q_s32(ptr2); in WebRtcIsacfix_FilterMaLoopNeon()
|
| /external/XNNPACK/src/qu8-requantization/ |
| D | q31-neon.c | 53 const int32x4_t x = vld1q_s32(input); in xnn_qu8_requantize_q31__neon()
54 const int32x4_t y = vld1q_s32(input + 4); in xnn_qu8_requantize_q31__neon()
55 const int32x4_t z = vld1q_s32(input + 8); in xnn_qu8_requantize_q31__neon()
56 const int32x4_t w = vld1q_s32(input + 12); in xnn_qu8_requantize_q31__neon()
|
| D | fp32-neon.c | 43 const int32x4_t x = vld1q_s32(input); in xnn_qu8_requantize_fp32__neon()
44 const int32x4_t y = vld1q_s32(input + 4); in xnn_qu8_requantize_fp32__neon()
45 const int32x4_t z = vld1q_s32(input + 8); in xnn_qu8_requantize_fp32__neon()
46 const int32x4_t w = vld1q_s32(input + 12); in xnn_qu8_requantize_fp32__neon()
|
| D | precise-neon.c | 49 const int32x4_t x = vld1q_s32(input); in xnn_qu8_requantize_precise__neon()
50 const int32x4_t y = vld1q_s32(input + 4); in xnn_qu8_requantize_precise__neon()
51 const int32x4_t z = vld1q_s32(input + 8); in xnn_qu8_requantize_precise__neon()
52 const int32x4_t w = vld1q_s32(input + 12); in xnn_qu8_requantize_precise__neon()
|
| /external/XNNPACK/src/qs8-requantization/ |
| D | q31-neon.c | 53 const int32x4_t x = vld1q_s32(input); in xnn_qs8_requantize_q31__neon()
54 const int32x4_t y = vld1q_s32(input + 4); in xnn_qs8_requantize_q31__neon()
55 const int32x4_t z = vld1q_s32(input + 8); in xnn_qs8_requantize_q31__neon()
56 const int32x4_t w = vld1q_s32(input + 12); in xnn_qs8_requantize_q31__neon()
|
| D | fp32-neon.c | 43 const int32x4_t x = vld1q_s32(input); in xnn_qs8_requantize_fp32__neon()
44 const int32x4_t y = vld1q_s32(input + 4); in xnn_qs8_requantize_fp32__neon()
45 const int32x4_t z = vld1q_s32(input + 8); in xnn_qs8_requantize_fp32__neon()
46 const int32x4_t w = vld1q_s32(input + 12); in xnn_qs8_requantize_fp32__neon()
|
| /external/webrtc/common_audio/signal_processing/ |
| D | min_max_operations_neon.c | 87 int32x4_t in32x4_0 = vld1q_s32(p_start); in WebRtcSpl_MaxAbsValueW32Neon()
89 int32x4_t in32x4_1 = vld1q_s32(p_start); in WebRtcSpl_MaxAbsValueW32Neon()
177 int32x4_t in32x4_0 = vld1q_s32(p_start); in WebRtcSpl_MaxValueW32Neon()
179 int32x4_t in32x4_1 = vld1q_s32(p_start); in WebRtcSpl_MaxValueW32Neon()
257 int32x4_t in32x4_0 = vld1q_s32(p_start); in WebRtcSpl_MinValueW32Neon()
259 int32x4_t in32x4_1 = vld1q_s32(p_start); in WebRtcSpl_MinValueW32Neon()
|
| /external/XNNPACK/src/qs8-gavgpool/gen/ |
| D | 7p7x-minmax-neon-c32-acc2.c | 218 int32x4_t vacc0123 = vld1q_s32(b); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
219 int32x4_t vacc4567 = vld1q_s32(b + 4); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
220 int32x4_t vacc89AB = vld1q_s32(b + 8); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
221 int32x4_t vaccCDEF = vld1q_s32(b + 12); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
222 int32x4_t vaccGHIJ = vld1q_s32(b + 16); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
223 int32x4_t vaccKLMN = vld1q_s32(b + 20); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
224 int32x4_t vaccOPQR = vld1q_s32(b + 24); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
225 int32x4_t vaccSTUV = vld1q_s32(b + 28); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
265 int32x4_t vacc0123 = vld1q_s32(b); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
266 int32x4_t vacc4567 = vld1q_s32(b + 4); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c32_acc2()
[all …]
|
| D | 7p7x-minmax-neon-c24-acc2.c | 188 int32x4_t vacc0123 = vld1q_s32(b); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
189 int32x4_t vacc4567 = vld1q_s32(b + 4); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
190 int32x4_t vacc89AB = vld1q_s32(b + 8); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
191 int32x4_t vaccCDEF = vld1q_s32(b + 12); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
192 int32x4_t vaccGHIJ = vld1q_s32(b + 16); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
193 int32x4_t vaccKLMN = vld1q_s32(b + 20); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
229 int32x4_t vacc0123 = vld1q_s32(b); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
230 int32x4_t vacc4567 = vld1q_s32(b + 4); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
323 int32x4_t vacc0123 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
324 int32x4_t vacc4567 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c24_acc2()
[all …]
|
| D | 7p7x-minmax-neon-c16-acc2.c | 129 int32x4_t vacc0123 = vld1q_s32(b); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
130 int32x4_t vacc4567 = vld1q_s32(b + 4); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
131 int32x4_t vacc89AB = vld1q_s32(b + 8); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
132 int32x4_t vaccCDEF = vld1q_s32(b + 12); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
213 int32x4_t vacc0123 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
214 int32x4_t vacc4567 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
215 int32x4_t vacc89AB = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
216 int32x4_t vaccCDEF = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
325 int32x4_t vacc0123 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
326 int32x4_t vacc4567 = vld1q_s32(buffer); buffer += 4; in xnn_qs8_gavgpool_minmax_ukernel_7p7x__neon_c16_acc2()
|
| /external/libgav1/libgav1/src/dsp/arm/ |
| D | inverse_transform_10bit_neon.cc | 86 x[i] = vld1q_s32(&src[i * stride + idx]); in LoadSrc()
87 x[i + 1] = vld1q_s32(&src[(i + 1) * stride + idx]); in LoadSrc()
88 x[i + 2] = vld1q_s32(&src[(i + 2) * stride + idx]); in LoadSrc()
89 x[i + 3] = vld1q_s32(&src[(i + 3) * stride + idx]); in LoadSrc()
229 const int32x4_t v_src = vld1q_s32(dst); in DctDcOnlyColumn()
235 const int32x4_t v_src = vld1q_s32(&dst[i]); in DctDcOnlyColumn()
1004 const int32x4_t kAdst4DcOnlyMultipliers = vld1q_s32(kAdst4DcOnlyMultiplier); in Adst4DcOnly()
1031 const int32x4_t v_src = vld1q_s32(&dst[i]); in Adst4DcOnlyColumn()
1198 const int32x4_t v_src = vld1q_s32(dst); in Adst8DcOnlyColumn()
1464 const int32x4_t v_src = vld1q_s32(dst); in Adst16DcOnlyColumn()
[all …]
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 1x16c4-minmax-neondot.c | 47 int32x4_t vacc0x0123 = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot()
48 int32x4_t vacc0x4567 = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot()
49 int32x4_t vacc0x89AB = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot()
50 int32x4_t vacc0xCDEF = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot()
|