[performance] cache library performance enhancements (updates go-structr => v0.2.0) (#2575)

* update go-structr => v0.2.0 * update readme * whoops, fix the link
2025-11-15 07:27:28 -06:00 · 2024-01-26 12:14:10 +00:00 · 2024-01-26 12:14:10 +00:00 · 07207e71e9
commit 07207e71e9
parent c946d02c1f
36 changed files with 4880 additions and 1379 deletions
--- a/vendor/github.com/zeebo/xxh3/hash64_seed.go
+++ b/vendor/github.com/zeebo/xxh3/hash64_seed.go
@ -0,0 +1,134 @@
+package xxh3
+
+import "math/bits"
+
+// HashSeed returns the hash of the byte slice with given seed.
+func HashSeed(b []byte, seed uint64) uint64 {
+	return hashAnySeed(*(*str)(ptr(&b)), seed)
+
+}
+
+// HashStringSeed returns the hash of the string slice with given seed.
+func HashStringSeed(s string, seed uint64) uint64 {
+	return hashAnySeed(*(*str)(ptr(&s)), seed)
+}
+
+func hashAnySeed(s str, seed uint64) (acc u64) {
+	p, l := s.p, s.l
+
+	switch {
+	case l <= 16:
+		switch {
+		case l > 8:
+			inputlo := readU64(p, 0) ^ (key64_024 ^ key64_032 + seed)
+			inputhi := readU64(p, ui(l)-8) ^ (key64_040 ^ key64_048 - seed)
+			folded := mulFold64(inputlo, inputhi)
+			return xxh3Avalanche(u64(l) + bits.ReverseBytes64(inputlo) + inputhi + folded)
+
+		case l > 3:
+			seed ^= u64(bits.ReverseBytes32(u32(seed))) << 32
+			input1 := readU32(p, 0)
+			input2 := readU32(p, ui(l)-4)
+			input64 := u64(input2) + u64(input1)<<32
+			keyed := input64 ^ (key64_008 ^ key64_016 - seed)
+			return rrmxmx(keyed, u64(l))
+
+		case l == 3: // 3
+			c12 := u64(readU16(p, 0))
+			c3 := u64(readU8(p, 2))
+			acc = c12<<16 + c3 + 3<<8
+
+		case l > 1: // 2
+			c12 := u64(readU16(p, 0))
+			acc = c12*(1<<24+1)>>8 + 2<<8
+
+		case l == 1: // 1
+			c1 := u64(readU8(p, 0))
+			acc = c1*(1<<24+1<<16+1) + 1<<8
+
+		default:
+			return xxhAvalancheSmall(seed ^ key64_056 ^ key64_064)
+		}
+
+		acc ^= u64(key32_000^key32_004) + seed
+		return xxhAvalancheSmall(acc)
+
+	case l <= 128:
+		acc = u64(l) * prime64_1
+
+		if l > 32 {
+			if l > 64 {
+				if l > 96 {
+					acc += mulFold64(readU64(p, 6*8)^(key64_096+seed), readU64(p, 7*8)^(key64_104-seed))
+					acc += mulFold64(readU64(p, ui(l)-8*8)^(key64_112+seed), readU64(p, ui(l)-7*8)^(key64_120-seed))
+				} // 96
+				acc += mulFold64(readU64(p, 4*8)^(key64_064+seed), readU64(p, 5*8)^(key64_072-seed))
+				acc += mulFold64(readU64(p, ui(l)-6*8)^(key64_080+seed), readU64(p, ui(l)-5*8)^(key64_088-seed))
+			} // 64
+			acc += mulFold64(readU64(p, 2*8)^(key64_032+seed), readU64(p, 3*8)^(key64_040-seed))
+			acc += mulFold64(readU64(p, ui(l)-4*8)^(key64_048+seed), readU64(p, ui(l)-3*8)^(key64_056-seed))
+		} // 32
+		acc += mulFold64(readU64(p, 0*8)^(key64_000+seed), readU64(p, 1*8)^(key64_008-seed))
+		acc += mulFold64(readU64(p, ui(l)-2*8)^(key64_016+seed), readU64(p, ui(l)-1*8)^(key64_024-seed))
+
+		return xxh3Avalanche(acc)
+
+	case l <= 240:
+		acc = u64(l) * prime64_1
+
+		acc += mulFold64(readU64(p, 0*16+0)^(key64_000+seed), readU64(p, 0*16+8)^(key64_008-seed))
+		acc += mulFold64(readU64(p, 1*16+0)^(key64_016+seed), readU64(p, 1*16+8)^(key64_024-seed))
+		acc += mulFold64(readU64(p, 2*16+0)^(key64_032+seed), readU64(p, 2*16+8)^(key64_040-seed))
+		acc += mulFold64(readU64(p, 3*16+0)^(key64_048+seed), readU64(p, 3*16+8)^(key64_056-seed))
+		acc += mulFold64(readU64(p, 4*16+0)^(key64_064+seed), readU64(p, 4*16+8)^(key64_072-seed))
+		acc += mulFold64(readU64(p, 5*16+0)^(key64_080+seed), readU64(p, 5*16+8)^(key64_088-seed))
+		acc += mulFold64(readU64(p, 6*16+0)^(key64_096+seed), readU64(p, 6*16+8)^(key64_104-seed))
+		acc += mulFold64(readU64(p, 7*16+0)^(key64_112+seed), readU64(p, 7*16+8)^(key64_120-seed))
+
+		// avalanche
+		acc = xxh3Avalanche(acc)
+
+		// trailing groups after 128
+		top := ui(l) &^ 15
+		for i := ui(8 * 16); i < top; i += 16 {
+			acc += mulFold64(readU64(p, i+0)^(readU64(key, i-125)+seed), readU64(p, i+8)^(readU64(key, i-117)-seed))
+		}
+
+		// last 16 bytes
+		acc += mulFold64(readU64(p, ui(l)-16)^(key64_119+seed), readU64(p, ui(l)-8)^(key64_127-seed))
+
+		return xxh3Avalanche(acc)
+
+	default:
+		acc = u64(l) * prime64_1
+
+		secret := key
+		if seed != 0 {
+			secret = ptr(&[secretSize]byte{})
+			initSecret(secret, seed)
+		}
+
+		accs := [8]u64{
+			prime32_3, prime64_1, prime64_2, prime64_3,
+			prime64_4, prime32_2, prime64_5, prime32_1,
+		}
+
+		if hasAVX512 && l >= avx512Switch {
+			accumAVX512(&accs, p, secret, u64(l))
+		} else if hasAVX2 {
+			accumAVX2(&accs, p, secret, u64(l))
+		} else if hasSSE2 {
+			accumSSE(&accs, p, secret, u64(l))
+		} else {
+			accumScalarSeed(&accs, p, secret, u64(l))
+		}
+
+		// merge accs
+		acc += mulFold64(accs[0]^readU64(secret, 11), accs[1]^readU64(secret, 19))
+		acc += mulFold64(accs[2]^readU64(secret, 27), accs[3]^readU64(secret, 35))
+		acc += mulFold64(accs[4]^readU64(secret, 43), accs[5]^readU64(secret, 51))
+		acc += mulFold64(accs[6]^readU64(secret, 59), accs[7]^readU64(secret, 67))
+
+		return xxh3Avalanche(acc)
+	}
+}