1 use super::Adler32Imp;
2
3 /// Resolves update implementation if CPU supports ssse3 instructions.
get_imp() -> Option<Adler32Imp>4 pub fn get_imp() -> Option<Adler32Imp> {
5 get_imp_inner()
6 }
7
8 #[inline]
9 #[cfg(all(feature = "std", any(target_arch = "x86", target_arch = "x86_64")))]
get_imp_inner() -> Option<Adler32Imp>10 fn get_imp_inner() -> Option<Adler32Imp> {
11 if std::is_x86_feature_detected!("ssse3") {
12 Some(imp::update)
13 } else {
14 None
15 }
16 }
17
18 #[inline]
19 #[cfg(all(
20 target_feature = "ssse3",
21 not(all(feature = "std", any(target_arch = "x86", target_arch = "x86_64")))
22 ))]
get_imp_inner() -> Option<Adler32Imp>23 fn get_imp_inner() -> Option<Adler32Imp> {
24 Some(imp::update)
25 }
26
27 #[inline]
28 #[cfg(all(
29 not(target_feature = "ssse3"),
30 not(all(feature = "std", any(target_arch = "x86", target_arch = "x86_64")))
31 ))]
get_imp_inner() -> Option<Adler32Imp>32 fn get_imp_inner() -> Option<Adler32Imp> {
33 None
34 }
35
36 #[cfg(all(
37 any(target_arch = "x86", target_arch = "x86_64"),
38 any(feature = "std", target_feature = "ssse3")
39 ))]
40 mod imp {
41 const MOD: u32 = 65521;
42 const NMAX: usize = 5552;
43 const BLOCK_SIZE: usize = 32;
44 const CHUNK_SIZE: usize = NMAX / BLOCK_SIZE * BLOCK_SIZE;
45
46 #[cfg(target_arch = "x86")]
47 use core::arch::x86::*;
48 #[cfg(target_arch = "x86_64")]
49 use core::arch::x86_64::*;
50
update(a: u16, b: u16, data: &[u8]) -> (u16, u16)51 pub fn update(a: u16, b: u16, data: &[u8]) -> (u16, u16) {
52 unsafe { update_imp(a, b, data) }
53 }
54
55 #[inline]
56 #[target_feature(enable = "ssse3")]
update_imp(a: u16, b: u16, data: &[u8]) -> (u16, u16)57 unsafe fn update_imp(a: u16, b: u16, data: &[u8]) -> (u16, u16) {
58 let mut a = a as u32;
59 let mut b = b as u32;
60
61 let chunks = data.chunks_exact(CHUNK_SIZE);
62 let remainder = chunks.remainder();
63 for chunk in chunks {
64 update_chunk_block(&mut a, &mut b, chunk);
65 }
66
67 update_block(&mut a, &mut b, remainder);
68
69 (a as u16, b as u16)
70 }
71
update_chunk_block(a: &mut u32, b: &mut u32, chunk: &[u8])72 unsafe fn update_chunk_block(a: &mut u32, b: &mut u32, chunk: &[u8]) {
73 debug_assert_eq!(
74 chunk.len(),
75 CHUNK_SIZE,
76 "Unexpected chunk size (expected {}, got {})",
77 CHUNK_SIZE,
78 chunk.len()
79 );
80
81 reduce_add_blocks(a, b, chunk);
82
83 *a %= MOD;
84 *b %= MOD;
85 }
86
update_block(a: &mut u32, b: &mut u32, chunk: &[u8])87 unsafe fn update_block(a: &mut u32, b: &mut u32, chunk: &[u8]) {
88 debug_assert!(
89 chunk.len() <= CHUNK_SIZE,
90 "Unexpected chunk size (expected <= {}, got {})",
91 CHUNK_SIZE,
92 chunk.len()
93 );
94
95 for byte in reduce_add_blocks(a, b, chunk) {
96 *a += *byte as u32;
97 *b += *a;
98 }
99
100 *a %= MOD;
101 *b %= MOD;
102 }
103
104 #[inline(always)]
reduce_add_blocks<'a>(a: &mut u32, b: &mut u32, chunk: &'a [u8]) -> &'a [u8]105 unsafe fn reduce_add_blocks<'a>(a: &mut u32, b: &mut u32, chunk: &'a [u8]) -> &'a [u8] {
106 if chunk.len() < BLOCK_SIZE {
107 return chunk;
108 }
109
110 let blocks = chunk.chunks_exact(BLOCK_SIZE);
111 let blocks_remainder = blocks.remainder();
112
113 let one_v = _mm_set1_epi16(1);
114 let zero_v = _mm_set1_epi16(0);
115 let weight_hi_v = get_weight_hi();
116 let weight_lo_v = get_weight_lo();
117
118 let mut p_v = _mm_set_epi32(0, 0, 0, (*a * blocks.len() as u32) as _);
119 let mut a_v = _mm_set_epi32(0, 0, 0, 0);
120 let mut b_v = _mm_set_epi32(0, 0, 0, *b as _);
121
122 for block in blocks {
123 let block_ptr = block.as_ptr() as *const _;
124 let left_v = _mm_loadu_si128(block_ptr);
125 let right_v = _mm_loadu_si128(block_ptr.add(1));
126
127 p_v = _mm_add_epi32(p_v, a_v);
128
129 a_v = _mm_add_epi32(a_v, _mm_sad_epu8(left_v, zero_v));
130 let mad = _mm_maddubs_epi16(left_v, weight_hi_v);
131 b_v = _mm_add_epi32(b_v, _mm_madd_epi16(mad, one_v));
132
133 a_v = _mm_add_epi32(a_v, _mm_sad_epu8(right_v, zero_v));
134 let mad = _mm_maddubs_epi16(right_v, weight_lo_v);
135 b_v = _mm_add_epi32(b_v, _mm_madd_epi16(mad, one_v));
136 }
137
138 b_v = _mm_add_epi32(b_v, _mm_slli_epi32(p_v, 5));
139
140 *a += reduce_add(a_v);
141 *b = reduce_add(b_v);
142
143 blocks_remainder
144 }
145
146 #[inline(always)]
reduce_add(v: __m128i) -> u32147 unsafe fn reduce_add(v: __m128i) -> u32 {
148 let hi = _mm_unpackhi_epi64(v, v);
149 let sum = _mm_add_epi32(hi, v);
150 let hi = _mm_shuffle_epi32(sum, crate::imp::_MM_SHUFFLE(2, 3, 0, 1));
151 let sum = _mm_add_epi32(sum, hi);
152
153 _mm_cvtsi128_si32(sum) as _
154 }
155
156 #[inline(always)]
get_weight_lo() -> __m128i157 unsafe fn get_weight_lo() -> __m128i {
158 _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
159 }
160
161 #[inline(always)]
get_weight_hi() -> __m128i162 unsafe fn get_weight_hi() -> __m128i {
163 _mm_set_epi8(
164 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32,
165 )
166 }
167 }
168
169 #[cfg(test)]
170 mod tests {
171 use rand::Rng;
172
173 #[test]
zeroes()174 fn zeroes() {
175 assert_sum_eq(&[]);
176 assert_sum_eq(&[0]);
177 assert_sum_eq(&[0, 0]);
178 assert_sum_eq(&[0; 100]);
179 assert_sum_eq(&[0; 1024]);
180 assert_sum_eq(&[0; 1024 * 1024]);
181 }
182
183 #[test]
ones()184 fn ones() {
185 assert_sum_eq(&[]);
186 assert_sum_eq(&[1]);
187 assert_sum_eq(&[1, 1]);
188 assert_sum_eq(&[1; 100]);
189 assert_sum_eq(&[1; 1024]);
190 assert_sum_eq(&[1; 1024 * 1024]);
191 }
192
193 #[test]
random()194 fn random() {
195 let mut random = [0; 1024 * 1024];
196 rand::thread_rng().fill(&mut random[..]);
197
198 assert_sum_eq(&random[..1]);
199 assert_sum_eq(&random[..100]);
200 assert_sum_eq(&random[..1024]);
201 assert_sum_eq(&random[..1024 * 1024]);
202 }
203
204 /// Example calculation from https://en.wikipedia.org/wiki/Adler-32.
205 #[test]
wiki()206 fn wiki() {
207 assert_sum_eq(b"Wikipedia");
208 }
209
assert_sum_eq(data: &[u8])210 fn assert_sum_eq(data: &[u8]) {
211 if let Some(update) = super::get_imp() {
212 let (a, b) = update(1, 0, data);
213 let left = u32::from(b) << 16 | u32::from(a);
214 let right = adler::adler32_slice(data);
215
216 assert_eq!(left, right, "len({})", data.len());
217 }
218 }
219 }
220