optimization: refactor: Introduce Uint256ExtraSipHasher to cache SipH…

…ash constant state

Previously, only k0 and k1 were stored, causing the constant xor operations to be recomputed in every call to `SipHashUint256Extra`.
This commit adds a dedicated `Uint256ExtraSipHasher` class that caches the initial state (v0-v3) and to perform the `SipHash` computation on a `uint256` (with an extra parameter), hiding the constant computation details from higher-level code and improving efficiency.

> cmake -B build -DBUILD_BENCH=ON -DCMAKE_BUILD_TYPE=Release && cmake --build build -j$(nproc) && build/src/bench/bench_bitcoin -filter='SaltedOutpointHasherBench' -min-time=10000

> C++ compiler .......................... AppleClang 16.0.0.16000026

|               ns/op |                op/s |    err% |     total | benchmark
|--------------------:|--------------------:|--------:|----------:|:----------
|               57.27 |       17,462,299.19 |    0.1% |     11.02 | `SaltedOutpointHasherBench_create_set`
|               11.24 |       88,997,888.48 |    0.3% |     11.04 | `SaltedOutpointHasherBench_hash`
|               13.91 |       71,902,014.20 |    0.2% |     11.01 | `SaltedOutpointHasherBench_match`
|               13.29 |       75,230,390.31 |    0.1% |     11.00 | `SaltedOutpointHasherBench_mismatch`

> C++ compiler .......................... GNU 13.3.0

?

compared to master:
SaltedOutpointHasherBenchmark_create_set - 17,462,299.19/17,065,922.04 - 2.3% faster
SaltedOutpointHasherBenchmark_hash - 88,997,888.48/83,576,684.83 - 6.4% faster
SaltedOutpointHasherBenchmark_match - 71,902,014.20/68,985,850.12 - 4.2% faster
SaltedOutpointHasherBenchmark_mismatch - 75,230,390.31/71,942,033.47 - 4.5% faster

src/crypto/siphash.cpp

@@ -132,17 +132,12 @@ uint64_t SipHashUint256(uint64_t k0, uint64_t k1, const uint256& val)
     return v0 ^ v1 ^ v2 ^ v3;
 }
 
-uint64_t SipHashUint256Extra(uint64_t k0, uint64_t k1, const uint256& val, uint32_t extra)
+/* Specialized implementation for efficiency */
+uint64_t Uint256ExtraSipHasher::operator()(const uint256& val, uint32_t extra) const noexcept
 {
-    /* Specialized implementation for efficiency */
+    uint64_t v0 = v[0], v1 = v[1], v2 = v[2], v3 = v[3];
     uint64_t d = val.GetUint64(0);
-
-    // TODO moved in next commit
-    uint64_t v0 = CSipHasher::C0 ^ k0;
-    uint64_t v1 = CSipHasher::C1 ^ k1;
-    uint64_t v2 = CSipHasher::C2 ^ k0;
-    uint64_t v3 = CSipHasher::C3 ^ k1 ^ d;
-
+    v3 ^= d;
     SIPROUND;
     SIPROUND;
     v0 ^= d;

src/crypto/siphash.h

@@ -48,6 +48,19 @@ class CSipHasher
  *      .Finalize()
  */
 uint64_t SipHashUint256(uint64_t k0, uint64_t k1, const uint256& val);
-uint64_t SipHashUint256Extra(uint64_t k0, uint64_t k1, const uint256& val, uint32_t extra);
+
+class Uint256ExtraSipHasher {
+    uint64_t v[4];
+
+public:
+    Uint256ExtraSipHasher(const uint64_t k0, const uint64_t k1) noexcept {
+        v[0] = CSipHasher::C0 ^ k0;
+        v[1] = CSipHasher::C1 ^ k1;
+        v[2] = CSipHasher::C2 ^ k0;
+        v[3] = CSipHasher::C3 ^ k1;
+    }
+
+    uint64_t operator()(const uint256& val, uint32_t extra) const noexcept;
+};
 
 #endif // BITCOIN_CRYPTO_SIPHASH_H

src/test/fuzz/integer.cpp

@@ -119,7 +119,7 @@ FUZZ_TARGET(integer, .init = initialize_integer)
     (void)MillisToTimeval(i64);
     (void)SighashToStr(uch);
     (void)SipHashUint256(u64, u64, u256);
-    (void)SipHashUint256Extra(u64, u64, u256, u32);
+    (void)Uint256ExtraSipHasher(u64, u64)(u256, u32);
     (void)ToLower(ch);
     (void)ToUpper(ch);
     {

src/test/hash_tests.cpp

@@ -130,7 +130,7 @@ BOOST_AUTO_TEST_CASE(siphash)
     ss << TX_WITH_WITNESS(tx);
     BOOST_CHECK_EQUAL(SipHashUint256(1, 2, ss.GetHash()), 0x79751e980c2a0a35ULL);
 
-    // Check consistency between CSipHasher and SipHashUint256[Extra].
+    // Check consistency between CSipHasher and SipHashUint256 and Uint256ExtraSipHasher.
     FastRandomContext ctx;
     for (int i = 0; i < 16; ++i) {
         uint64_t k0 = ctx.rand64();
@@ -144,7 +144,7 @@ BOOST_AUTO_TEST_CASE(siphash)
         CSipHasher sip288 = sip256;
         sip288.Write(nb);
         BOOST_CHECK_EQUAL(SipHashUint256(k0, k1, x), sip256.Finalize());
-        BOOST_CHECK_EQUAL(SipHashUint256Extra(k0, k1, x, n), sip288.Finalize()); // TODO modified in follow-up commit
+        BOOST_CHECK_EQUAL(Uint256ExtraSipHasher(k0, k1)(x, n), sip288.Finalize());
     }
 }
 

src/util/hasher.cpp

@@ -11,9 +11,9 @@ SaltedTxidHasher::SaltedTxidHasher() :
     k0{FastRandomContext().rand64()},
     k1{FastRandomContext().rand64()} {}
 
-SaltedOutpointHasher::SaltedOutpointHasher(bool deterministic) :
-    k0{deterministic ? 0x8e819f2607a18de6 : FastRandomContext().rand64()},
-    k1{deterministic ? 0xf4020d2e3983b0eb : FastRandomContext().rand64()}
+SaltedOutpointHasher::SaltedOutpointHasher(bool deterministic) : hasher{
+    deterministic ? 0x8e819f2607a18de6 : FastRandomContext().rand64(),
+    deterministic ? 0xf4020d2e3983b0eb : FastRandomContext().rand64()}
 {}
 
 SaltedSipHasher::SaltedSipHasher() :

src/util/hasher.h

@@ -30,12 +30,10 @@ class SaltedTxidHasher
 
 class SaltedOutpointHasher
 {
-private:
-    /** Salt */
-    const uint64_t k0, k1;
+    const Uint256ExtraSipHasher hasher;
 
 public:
-    SaltedOutpointHasher(bool deterministic = false);
+    explicit SaltedOutpointHasher(bool deterministic = false);
 
     /**
      * Having the hash noexcept allows libstdc++'s unordered_map to recalculate
@@ -47,7 +45,7 @@ class SaltedOutpointHasher
      * @see https://gcc.gnu.org/onlinedocs/gcc-13.2.0/libstdc++/manual/manual/unordered_associative.html
      */
     size_t operator()(const COutPoint& id) const noexcept {
-        return SipHashUint256Extra(k0, k1, id.hash, id.n);
+        return hasher(id.hash, id.n);
     }
 };