cor_h_vec_neon.s revision f3664ae9369a861ffbc2354e8e93e48983802062
1@/* 2@ ** Copyright 2003-2010, VisualOn, Inc. 3@ ** 4@ ** Licensed under the Apache License, Version 2.0 (the "License"); 5@ ** you may not use this file except in compliance with the License. 6@ ** You may obtain a copy of the License at 7@ ** 8@ ** http://www.apache.org/licenses/LICENSE-2.0 9@ ** 10@ ** Unless required by applicable law or agreed to in writing, software 11@ ** distributed under the License is distributed on an "AS IS" BASIS, 12@ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13@ ** See the License for the specific language governing permissions and 14@ ** limitations under the License. 15@ */ 16@ 17@static void cor_h_vec_012( 18@ Word16 h[], /* (i) scaled impulse response */ 19@ Word16 vec[], /* (i) scaled vector (/8) to correlate with h[] */ 20@ Word16 track, /* (i) track to use */ 21@ Word16 sign[], /* (i) sign vector */ 22@ Word16 rrixix[][NB_POS], /* (i) correlation of h[x] with h[x] */ 23@ Word16 cor_1[], /* (o) result of correlation (NB_POS elements) */ 24@ Word16 cor_2[] /* (o) result of correlation (NB_POS elements) */ 25@) 26@r0 ---- h[] 27@r1 ---- vec[] 28@r2 ---- track 29@r3 ---- sign[] 30@r4 ---- rrixix[][NB_POS] 31@r5 ---- cor_1[] 32@r6 ---- cor_2[] 33 34 .section .text 35 .global cor_h_vec_012_asm 36 37cor_h_vec_012_asm: 38 39 STMFD r13!, {r4 - r12, r14} 40 LDR r4, [r13, #40] @load rrixix[][NB_POS] 41 ADD r7, r4, r2, LSL #5 @r7 --- p0 = rrixix[track] 42 MOV r4, #0 @i=0 43 44 @r0 --- h[], r1 --- vec[], r2 --- pos 45 @r3 --- sign[], r4 --- i, r7 --- p0 46 47LOOPi: 48 MOV r5, #0 @L_sum1 = 0 49 MOV r6, #0 @L_sum2 = 0 50 ADD r9, r1, r2, LSL #1 @p2 = &vec[pos] 51 MOV r10, r0 @p1 = h 52 RSB r11, r2, #62 @j=62-pos 53 54LOOPj1: 55 LDRSH r12, [r10], #2 56 LDRSH r8, [r9], #2 57 LDRSH r14, [r9] 58 SUBS r11, r11, #1 59 MLA r5, r12, r8, r5 60 MLA r6, r12, r14, r6 61 BGE LOOPj1 62 63 LDRSH r12, [r10], #2 @*p1++ 64 MOV r6, r6, LSL #2 @L_sum2 = (L_sum2 << 2) 65 MLA r5, r12, r14, r5 66 MOV r14, #0x8000 67 MOV r5, r5, LSL #2 @L_sum1 = (L_sum1 << 2) 68 ADD r10, r6, r14 69 ADD r9, r5, r14 70 MOV r5, r9, ASR #16 71 MOV r6, r10, ASR #16 72 ADD r9, r3, r2, LSL #1 @address of sign[pos] 73 ADD r8, r7, #32 74 LDRSH r10, [r9], #2 @sign[pos] 75 LDRSH r11, [r9] @sign[pos + 1] 76 MUL r12, r5, r10 77 MUL r14, r6, r11 78 MOV r5, r12, ASR #15 79 MOV r6, r14, ASR #15 80 LDR r9, [r13, #44] 81 LDR r12, [r13, #48] 82 LDRSH r10, [r7], #2 @*p0++ 83 LDRSH r11, [r8] @*p3++ 84 ADD r9, r9, r4, LSL #1 85 ADD r12, r12, r4, LSL #1 86 ADD r5, r5, r10 87 ADD r6, r6, r11 88 STRH r5, [r9] 89 STRH r6, [r12] 90 91 ADD r2, r2, #4 92 93 MOV r5, #0 @L_sum1 = 0 94 MOV r6, #0 @L_sum2 = 0 95 ADD r9, r1, r2, LSL #1 @p2 = &vec[pos] 96 MOV r10, r0 @p1 = h 97 RSB r11, r2, #62 @j=62-pos 98 ADD r4, r4, #1 @i++ 99 100LOOPj2: 101 LDRSH r12, [r10], #2 102 LDRSH r8, [r9], #2 103 LDRSH r14, [r9] 104 SUBS r11, r11, #1 105 MLA r5, r12, r8, r5 106 MLA r6, r12, r14, r6 107 BGE LOOPj2 108 109 LDRSH r12, [r10], #2 @*p1++ 110 MOV r6, r6, LSL #2 @L_sum2 = (L_sum2 << 2) 111 MLA r5, r12, r14, r5 112 MOV r14, #0x8000 113 MOV r5, r5, LSL #2 @L_sum1 = (L_sum1 << 2) 114 ADD r10, r6, r14 115 ADD r9, r5, r14 116 117 MOV r5, r9, ASR #16 118 MOV r6, r10, ASR #16 119 ADD r9, r3, r2, LSL #1 @address of sign[pos] 120 ADD r8, r7, #32 121 LDRSH r10, [r9], #2 @sign[pos] 122 LDRSH r11, [r9] @sign[pos + 1] 123 MUL r12, r5, r10 124 MUL r14, r6, r11 125 MOV r5, r12, ASR #15 126 MOV r6, r14, ASR #15 127 LDR r9, [r13, #44] 128 LDR r12, [r13, #48] 129 LDRSH r10, [r7], #2 @*p0++ 130 LDRSH r11, [r8] @*p3++ 131 ADD r9, r9, r4, LSL #1 132 ADD r12, r12, r4, LSL #1 133 ADD r5, r5, r10 134 ADD r6, r6, r11 135 STRH r5, [r9] 136 STRH r6, [r12] 137 ADD r4, r4, #1 @i+1 138 ADD r2, r2, #4 @pos += STEP 139 CMP r4, #16 140 141 BLT LOOPi 142 143the_end: 144 LDMFD r13!, {r4 - r12, r15} 145 146 .END 147 148 149 150 151 152