| /external/libvpx/libvpx/vp8/encoder/arm/neon/ |
| D | shortfdct_neon.c | 18 int32x4_t q9s32, q10s32, q11s32, q12s32; in vp8_short_fdct4x4_neon() local
24 q9s32 = vdupq_n_s32(14500); in vp8_short_fdct4x4_neon()
59 q9s32 = vmlal_s16(q9s32, d7s16, d16s16); in vp8_short_fdct4x4_neon()
61 q9s32 = vmlal_s16(q9s32, d6s16, d17s16); in vp8_short_fdct4x4_neon()
64 d1s16 = vshrn_n_s32(q9s32, 12); in vp8_short_fdct4x4_neon()
118 int32x4_t q9s32, q10s32, q11s32, q12s32; in vp8_short_fdct8x4_neon() local
124 q9s32 = vdupq_n_s32(14500); in vp8_short_fdct8x4_neon()
159 q11s32 = q9s32; in vp8_short_fdct8x4_neon()
167 q9s32 = vmlal_s16(q9s32, d28s16, d16s16); in vp8_short_fdct8x4_neon()
172 q9s32 = vmlal_s16(q9s32, d26s16, d17s16); in vp8_short_fdct8x4_neon()
[all …]
|
| D | vp8_shortwalsh4x4_neon.c | 25 int32x4_t q9s32, q10s32, q11s32, q15s32; in vp8_short_walsh4x4_neon() local
78 q9s32 = vaddl_s16(v2tmp1.val[1], v2tmp0.val[1]); in vp8_short_walsh4x4_neon()
82 q0s32 = vaddq_s32(q8s32, q9s32); in vp8_short_walsh4x4_neon()
85 q3s32 = vsubq_s32(q8s32, q9s32); in vp8_short_walsh4x4_neon()
93 q9s32 = vreinterpretq_s32_u32(q9u32); in vp8_short_walsh4x4_neon()
98 q1s32 = vsubq_s32(q1s32, q9s32); in vp8_short_walsh4x4_neon()
103 q9s32 = vaddq_s32(q1s32, q15s32); in vp8_short_walsh4x4_neon()
108 d1s16 = vshrn_n_s32(q9s32, 3); in vp8_short_walsh4x4_neon()
|
| /external/libvpx/libvpx/vpx_dsp/arm/ |
| D | variance_neon.c | 267 int32x4_t q7s32, q8s32, q9s32, q10s32; in vpx_mse16x16_neon() local
273 q9s32 = vdupq_n_s32(0); in vpx_mse16x16_neon()
298 q9s32 = vmlal_s16(q9s32, d24s16, d24s16); in vpx_mse16x16_neon()
308 q9s32 = vmlal_s16(q9s32, d28s16, d28s16); in vpx_mse16x16_neon()
313 q9s32 = vaddq_s32(q9s32, q10s32); in vpx_mse16x16_neon()
314 q10s32 = vaddq_s32(q7s32, q9s32); in vpx_mse16x16_neon()
330 int32x4_t q7s32, q8s32, q9s32, q10s32; in vpx_get4x4sse_cs_neon() local
363 q9s32 = vmull_s16(d26s16, d26s16); in vpx_get4x4sse_cs_neon()
367 q9s32 = vaddq_s32(q9s32, q10s32); in vpx_get4x4sse_cs_neon()
368 q9s32 = vaddq_s32(q7s32, q9s32); in vpx_get4x4sse_cs_neon()
[all …]
|
| /external/libvpx/libvpx/vp9/common/arm/neon/ |
| D | vp9_iht8x8_add_neon.c | 44 int32x4_t q2s32, q3s32, q5s32, q6s32, q8s32, q9s32; in IDCT8x8_1D() local
88 q9s32 = vmull_s16(d26s16, d3s16); in IDCT8x8_1D()
93 q9s32 = vmlal_s16(q9s32, d22s16, d2s16); in IDCT8x8_1D()
98 d12s16 = vrshrn_n_s32(q9s32, 14); in IDCT8x8_1D()
158 q9s32 = vmull_s16(d28s16, d16s16); in IDCT8x8_1D()
163 q9s32 = vmlsl_s16(q9s32, d26s16, d16s16); in IDCT8x8_1D()
168 d10s16 = vrshrn_n_s32(q9s32, 14); in IDCT8x8_1D()
195 int32x4_t q9s32, q10s32, q11s32, q12s32, q13s32, q14s32, q15s32; in IADST8X8_1D() local
284 q9s32 = vmull_s16(d19s16, d31s16); in IADST8X8_1D()
287 q9s32 = vmlsl_s16(q9s32, d29s16, d30s16); in IADST8X8_1D()
[all …]
|
| D | vp9_iht4x4_add_neon.c | 27 int32x4_t q8s32, q9s32; in TRANSPOSE4X4() local
35 q9s32 = vreinterpretq_s32_s16(vcombine_s16(d1x2s16.val[0], d1x2s16.val[1])); in TRANSPOSE4X4()
36 q0x2s32 = vtrnq_s32(q8s32, q9s32); in TRANSPOSE4X4()
96 int32x4_t q8s32, q9s32, q10s32, q11s32, q12s32, q13s32, q14s32, q15s32; in IADST4x4_1D() local
114 q9s32 = vmull_s16(*d5s16, d19s16); in IADST4x4_1D()
120 q11s32 = vsubq_s32(q11s32, q9s32); in IADST4x4_1D()
|