2 * ARM NEON optimised integer operations
3 * Copyright (c) 2009 Konstantin Shishkov
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 #include "libavutil/arm/asm.S"
24 function ff_scalarproduct_int16_neon, export=1
29 1: vld1.16 {d16-d17}, [r0]!
30 vld1.16 {d20-d21}, [r1,:128]!
31 vmlal.s16 q0, d16, d20
32 vld1.16 {d18-d19}, [r0]!
33 vmlal.s16 q1, d17, d21
34 vld1.16 {d22-d23}, [r1,:128]!
35 vmlal.s16 q2, d18, d22
36 vmlal.s16 q3, d19, d23
44 vpadd.s32 d0, d16, d17
45 vpadd.s32 d1, d18, d19
52 @ scalarproduct_and_madd_int16(/*aligned*/v0,v1,v2,order,mul)
53 function ff_scalarproduct_and_madd_int16_neon, export=1
54 vld1.16 {d28[],d29[]}, [sp]
61 1: vld1.16 {d16-d17}, [r0,:128]!
62 vld1.16 {d18-d19}, [r1]!
63 vld1.16 {d20-d21}, [r2]!
64 vld1.16 {d22-d23}, [r0,:128]!
65 vld1.16 {d24-d25}, [r1]!
66 vld1.16 {d26-d27}, [r2]!
67 vmul.s16 q10, q10, q14
68 vmul.s16 q13, q13, q14
69 vmlal.s16 q0, d16, d18
70 vmlal.s16 q1, d17, d19
72 vadd.s16 q13, q11, q13
73 vmlal.s16 q2, d22, d24
74 vmlal.s16 q3, d23, d25
75 vst1.16 {q10}, [r12,:128]!
77 vst1.16 {q13}, [r12,:128]!
84 vpadd.s32 d0, d16, d17
85 vpadd.s32 d1, d18, d19