1 /*
2 * Copyright (c) 2005
3 * Eric Anholt. All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 *
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24 * SUCH DAMAGE.
25 */
26
27 #ifdef HAVE_CONFIG_H
28 #include "config.h"
29 #endif
30 #include <liboil/liboilclasses.h>
31 #include <liboil/liboilfunction.h>
32 #include <mmintrin.h>
33
34 #ifdef ENABLE_BROKEN_IMPLS
35 union m64_int {
36 __m64 m64;
37 uint64_t ull;
38 };
39
40 static const struct _MMXData {
41 union m64_int mmx_4x0080;
42 } c = {
43 .mmx_4x0080.ull = 0x0080008000800080ULL,
44 };
45
46 #define MC(x) (c.mmx_##x.m64)
47
48 static void
recon8x8_intra_mmx(uint8_t * dest,int ds,int16_t * change)49 recon8x8_intra_mmx (uint8_t *dest, int ds, int16_t *change)
50 {
51 int i;
52 __m64 offset = MC(4x0080);
53
54 for (i = 8; i; i--) {
55 __m64 mm0, mm1, c0, c1;
56 c0 = ((__m64 *)change)[0];
57 c1 = ((__m64 *)change)[1];
58 mm0 = _mm_adds_pi16(c0, offset);
59 mm1 = _mm_adds_pi16(c1, offset);
60 *(__m64 *)dest = _mm_packs_pu16(mm0, mm1);
61
62 dest += ds;
63 change += 8;
64 }
65 _mm_empty();
66 }
67 OIL_DEFINE_IMPL_FULL (recon8x8_intra_mmx, recon8x8_intra, OIL_IMPL_FLAG_MMX);
68 #endif
69
70 static void
recon8x8_inter_mmx(uint8_t * dest,int ds,uint8_t * src,int ss,int16_t * change,int dss)71 recon8x8_inter_mmx (uint8_t *dest, int ds, uint8_t *src, int ss,
72 int16_t *change, int dss)
73 {
74 int i;
75
76 for (i = 8; i; i--) {
77 __m64 mm0, mm1, c0, c1;
78 c0 = ((__m64 *)change)[0];
79 c1 = ((__m64 *)change)[1];
80 mm0 = _mm_unpacklo_pi8(*(__m64 *)src, _mm_setzero_si64());
81 mm1 = _mm_unpackhi_pi8(*(__m64 *)src, _mm_setzero_si64());
82 mm0 = _mm_adds_pi16(mm0, c0);
83 mm1 = _mm_adds_pi16(mm1, c1);
84 *(__m64 *)dest = _mm_packs_pu16(mm0, mm1);
85
86 change += 8;
87 dest += ds;
88 src += ss;
89 }
90 _mm_empty();
91 }
92 OIL_DEFINE_IMPL_FULL (recon8x8_inter_mmx, recon8x8_inter, OIL_IMPL_FLAG_MMX);
93
94 static void
recon8x8_inter2_mmx(uint8_t * dest,int ds,uint8_t * s1,int ss1,uint8_t * s2,int ss2,int16_t * change)95 recon8x8_inter2_mmx (uint8_t *dest, int ds, uint8_t *s1, int ss1, uint8_t *s2,
96 int ss2, int16_t *change)
97 {
98 int i;
99
100 for (i = 8; i; i--) {
101 __m64 mm0, mm1, c0, c1;
102 mm0 = _mm_adds_pu16(
103 _mm_unpacklo_pi8(*(__m64 *)s1, _mm_setzero_si64()),
104 _mm_unpacklo_pi8(*(__m64 *)s2, _mm_setzero_si64()));
105 mm1 = _mm_adds_pu16(
106 _mm_unpackhi_pi8(*(__m64 *)s1, _mm_setzero_si64()),
107 _mm_unpackhi_pi8(*(__m64 *)s2, _mm_setzero_si64()));
108 c0 = ((__m64 *)change)[0];
109 c1 = ((__m64 *)change)[1];
110 mm0 = _mm_srli_pi16(mm0, 1);
111 mm1 = _mm_srli_pi16(mm1, 1);
112 mm0 = _mm_adds_pi16(mm0, c0);
113 mm1 = _mm_adds_pi16(mm1, c1);
114 *(__m64 *)dest = _mm_packs_pu16(mm0, mm1);
115
116 change += 8;
117 dest += ds;
118 s1 += ss1;
119 s2 += ss2;
120 }
121 _mm_empty();
122 }
123 OIL_DEFINE_IMPL_FULL (recon8x8_inter2_mmx, recon8x8_inter2, OIL_IMPL_FLAG_MMX);
124