/external/libvpx/libvpx/vpx_dsp/arm/ |
D | idct4x4_add_neon.c | 20 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16; in vpx_idct4x4_16_add_neon() local 37 d17s16 = vget_high_s16(q8s16); in vpx_idct4x4_16_add_neon() 41 d0x2s16 = vtrn_s16(d16s16, d17s16); in vpx_idct4x4_16_add_neon() 52 d17s16 = vget_high_s16(vreinterpretq_s16_s32(q0x2s32.val[0])); in vpx_idct4x4_16_add_neon() 62 q15s32 = vmull_s16(d17s16, d22s16); in vpx_idct4x4_16_add_neon() 63 q1s32 = vmull_s16(d17s16, d20s16); in vpx_idct4x4_16_add_neon() 82 d17s16 = vget_high_s16(q8s16); in vpx_idct4x4_16_add_neon() 86 d0x2s16 = vtrn_s16(d16s16, d17s16); in vpx_idct4x4_16_add_neon() 94 d17s16 = vget_high_s16(vreinterpretq_s16_s32(q0x2s32.val[0])); in vpx_idct4x4_16_add_neon() 103 q15s32 = vmull_s16(d17s16, d22s16); in vpx_idct4x4_16_add_neon() [all …]
|
D | vpx_convolve8_neon.c | 63 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d22s16, d23s16; in vpx_convolve8_horiz_neon() local 160 d17s16 = vreinterpret_s16_u16(vget_high_u16(q8u16)); in vpx_convolve8_horiz_neon() 169 q1s32 = MULTIPLY_BY_Q0(d16s16, d17s16, d20s16, d22s16, in vpx_convolve8_horiz_neon() 171 q2s32 = MULTIPLY_BY_Q0(d17s16, d20s16, d22s16, d18s16, in vpx_convolve8_horiz_neon() 236 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16; in vpx_convolve8_vert_neon() local 287 d17s16 = vreinterpret_s16_u16(vget_high_u16(q8u16)); in vpx_convolve8_vert_neon() 297 q1s32 = MULTIPLY_BY_Q0(d16s16, d17s16, d18s16, d19s16, in vpx_convolve8_vert_neon() 301 q2s32 = MULTIPLY_BY_Q0(d17s16, d18s16, d19s16, d20s16, in vpx_convolve8_vert_neon()
|
D | vpx_convolve8_avg_neon.c | 63 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d22s16, d23s16; in vpx_convolve8_avg_horiz_neon() local 167 d17s16 = vreinterpret_s16_u16(vget_high_u16(q8u16)); in vpx_convolve8_avg_horiz_neon() 176 q1s32 = MULTIPLY_BY_Q0(d16s16, d17s16, d20s16, d22s16, in vpx_convolve8_avg_horiz_neon() 178 q2s32 = MULTIPLY_BY_Q0(d17s16, d20s16, d22s16, d18s16, in vpx_convolve8_avg_horiz_neon() 252 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16; in vpx_convolve8_avg_vert_neon() local 312 d17s16 = vreinterpret_s16_u16(vget_high_u16(q8u16)); in vpx_convolve8_avg_vert_neon() 322 q1s32 = MULTIPLY_BY_Q0(d16s16, d17s16, d18s16, d19s16, in vpx_convolve8_avg_vert_neon() 326 q2s32 = MULTIPLY_BY_Q0(d17s16, d18s16, d19s16, d20s16, in vpx_convolve8_avg_vert_neon()
|
D | idct8x8_add_neon.c | 25 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in TRANSPOSE8X8() local 31 d17s16 = vget_high_s16(*q8s16); in TRANSPOSE8X8() 51 *q12s16 = vcombine_s16(d17s16, d25s16); in TRANSPOSE8X8() 96 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in IDCT8x8_1D() local 108 d17s16 = vget_high_s16(*q8s16); in IDCT8x8_1D() 161 q3s32 = vmull_s16(d17s16, d0s16); in IDCT8x8_1D() 163 q15s32 = vmull_s16(d17s16, d0s16); in IDCT8x8_1D()
|
D | idct16x16_add_neon.c | 25 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in TRANSPOSE8X8() local 31 d17s16 = vget_high_s16(*q8s16); in TRANSPOSE8X8() 51 *q12s16 = vcombine_s16(d17s16, d25s16); in TRANSPOSE8X8() 91 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in vpx_idct16x16_256_add_neon_pass1() local 129 d17s16 = vget_high_s16(q8s16); in vpx_idct16x16_256_add_neon_pass1() 190 q11s32 = vmull_s16(d17s16, d30s16); in vpx_idct16x16_256_add_neon_pass1() 203 d17s16 = vqrshrn_n_s32(q12s32, 14); in vpx_idct16x16_256_add_neon_pass1() 206 q8s16 = vcombine_s16(d16s16, d17s16); in vpx_idct16x16_256_add_neon_pass1() 334 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in vpx_idct16x16_256_add_neon_pass2() local 374 d17s16 = vget_high_s16(q8s16); in vpx_idct16x16_256_add_neon_pass2() [all …]
|
D | idct32x32_add_neon.c | 174 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in idct32_transpose_pair() local 199 d17s16 = vget_high_s16(q8s16); in idct32_transpose_pair() 219 q12s16 = vcombine_s16(d17s16, d25s16); in idct32_transpose_pair()
|
/external/libvpx/libvpx/vp8/encoder/arm/neon/ |
D | shortfdct_neon.c | 18 int16x4_t d16s16, d17s16, d26s16, dEmptys16; in vp8_short_fdct4x4_neon() local 26 d17s16 = vdup_n_s16(2217); in vp8_short_fdct4x4_neon() 65 q10s32 = vmlal_s16(q10s32, d7s16, d17s16); in vp8_short_fdct4x4_neon() 66 q9s32 = vmlal_s16(q9s32, d6s16, d17s16); in vp8_short_fdct4x4_neon() 94 q12s32 = vmlal_s16(q12s32, d7s16, d17s16); in vp8_short_fdct4x4_neon() 102 q11s32 = vmlal_s16(q11s32, d6s16, d17s16); in vp8_short_fdct4x4_neon() 123 int16x4_t d16s16, d17s16, d26s16, d27s16, d28s16, d29s16; in vp8_short_fdct8x4_neon() local 133 d17s16 = vdup_n_s16(2217); in vp8_short_fdct8x4_neon() 178 q10s32 = vmlal_s16(q10s32, d28s16, d17s16); in vp8_short_fdct8x4_neon() 180 q12s32 = vmlal_s16(q12s32, d29s16, d17s16); in vp8_short_fdct8x4_neon() [all …]
|
/external/libvpx/libvpx/vp9/common/arm/neon/ |
D | vp9_iht4x4_add_neon.c | 73 int16x4_t d16s16, d17s16, d18s16, d19s16, d23s16, d24s16; in IDCT4x4_1D() local 79 d17s16 = vget_high_s16(*q8s16); in IDCT4x4_1D() 86 q15s32 = vmull_s16(d17s16, *d2s16); in IDCT4x4_1D() 87 q10s32 = vmull_s16(d17s16, *d0s16); in IDCT4x4_1D() 114 int16x4_t d6s16, d16s16, d17s16, d18s16, d19s16; in IADST4x4_1D() local 120 d17s16 = vget_high_s16(*q8s16); in IADST4x4_1D() 126 q12s32 = vmull_s16(d6s16, d17s16); in IADST4x4_1D() 148 d17s16 = vqrshrn_n_s32(q14s32, 14); in IADST4x4_1D() 152 *q8s16 = vcombine_s16(d16s16, d17s16); in IADST4x4_1D()
|
D | vp9_iht8x8_add_neon.c | 43 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in TRANSPOSE8X8() local 49 d17s16 = vget_high_s16(*q8s16); in TRANSPOSE8X8() 69 *q12s16 = vcombine_s16(d17s16, d25s16); in TRANSPOSE8X8() 114 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in IDCT8x8_1D() local 126 d17s16 = vget_high_s16(*q8s16); in IDCT8x8_1D() 179 q3s32 = vmull_s16(d17s16, d0s16); in IDCT8x8_1D() 181 q15s32 = vmull_s16(d17s16, d0s16); in IDCT8x8_1D() 270 int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16, d23s16; in IADST8X8_1D() local 277 d17s16 = vget_high_s16(*q8s16); in IADST8X8_1D() 305 q2s32 = vmlal_s16(q2s32, d17s16, d15s16); in IADST8X8_1D() [all …]
|