From f47772f6f4efcd46bc571882e5d971f7301dc7ec Mon Sep 17 00:00:00 2001
From: Peter Dimov <pdimov@gmail.com>
Date: Fri, 16 Sep 2022 21:25:20 +0300
Subject: [PATCH] Implement faster hash_range when value_type is char

---
 .../container_hash/detail/hash_range.hpp      | 158 ++++++++++++++++++
 include/boost/container_hash/hash.hpp         |   6 +-
 2 files changed, 160 insertions(+), 4 deletions(-)
 create mode 100644 include/boost/container_hash/detail/hash_range.hpp

diff --git a/include/boost/container_hash/detail/hash_range.hpp b/include/boost/container_hash/detail/hash_range.hpp
new file mode 100644
index 0000000..7450e1e
--- /dev/null
+++ b/include/boost/container_hash/detail/hash_range.hpp
@@ -0,0 +1,158 @@
+// Copyright 2022 Peter Dimov
+// Distributed under the Boost Software License, Version 1.0.
+// https://www.boost.org/LICENSE_1_0.txt
+
+#ifndef BOOST_HASH_DETAIL_HASH_RANGE_HPP
+#define BOOST_HASH_DETAIL_HASH_RANGE_HPP
+
+#include <boost/container_hash/hash_fwd.hpp>
+#include <boost/type_traits/integral_constant.hpp>
+#include <boost/type_traits/enable_if.hpp>
+#include <boost/cstdint.hpp>
+#include <cstddef>
+#include <climits>
+#include <iterator>
+
+namespace boost
+{
+namespace hash_detail
+{
+
+template<class T> struct is_char_type: public boost::false_type {};
+
+#if CHAR_BIT == 8
+
+template<> struct is_char_type<char>: public boost::true_type {};
+template<> struct is_char_type<signed char>: public boost::true_type {};
+template<> struct is_char_type<unsigned char>: public boost::true_type {};
+
+#endif
+
+template<class It>
+inline typename boost::enable_if_< !is_char_type<typename std::iterator_traits<It>::value_type>::value >::type
+    hash_range( std::size_t& seed, It first, It last )
+{
+    for( ; first != last; ++first )
+    {
+        hash_combine<typename std::iterator_traits<It>::value_type>( seed, *first );
+    }
+}
+
+template<class It>
+inline typename boost::enable_if_<
+    is_char_type<typename std::iterator_traits<It>::value_type>::value &&
+    is_same<typename std::iterator_traits<It>::iterator_category, std::random_access_iterator_tag>::value
+>::type
+    hash_range( std::size_t& seed, It first, It last )
+{
+    std::size_t n = static_cast<std::size_t>( last - first );
+
+    for( ; n >= 4; first += 4, n -= 4 )
+    {
+        // clang 5+, gcc 5+ figure out this pattern and use a single mov on x86
+        // gcc on s390x and power BE even knows how to use load-reverse
+
+        boost::uint32_t w =
+            static_cast<boost::uint32_t>( static_cast<unsigned char>( first[0] ) ) |
+            static_cast<boost::uint32_t>( static_cast<unsigned char>( first[1] ) ) <<  8 |
+            static_cast<boost::uint32_t>( static_cast<unsigned char>( first[2] ) ) << 16 |
+            static_cast<boost::uint32_t>( static_cast<unsigned char>( first[3] ) ) << 24;
+
+        hash_combine( seed, w );
+    }
+
+    {
+        // add a trailing suffix byte of 0x01 because otherwise sequences of
+        // trailing zeroes are indistinguishable from end of string
+
+        boost::uint32_t w = 0x01u;
+
+        switch( n )
+        {
+        case 1:
+
+            w =
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[0] ) ) |
+                0x0100u;
+
+            break;
+
+        case 2:
+
+            w =
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[0] ) ) |
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[1] ) ) <<  8 |
+                0x010000u;
+
+            break;
+
+        case 3:
+
+            w =
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[0] ) ) |
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[1] ) ) <<  8 |
+                static_cast<boost::uint32_t>( static_cast<unsigned char>( first[2] ) ) << 16 |
+                0x01000000u;
+
+            break;
+        }
+
+        hash_combine( seed, w );
+    }
+}
+
+template<class It>
+inline typename boost::enable_if_<
+    is_char_type<typename std::iterator_traits<It>::value_type>::value &&
+    !is_same<typename std::iterator_traits<It>::iterator_category, std::random_access_iterator_tag>::value
+>::type
+    hash_range( std::size_t& seed, It first, It last )
+{
+    for( ;; )
+    {
+        boost::uint32_t w = 0;
+
+        if( first == last )
+        {
+            hash_combine( seed, w | 0x01u );
+            return;
+        }
+
+        w |= static_cast<boost::uint32_t>( static_cast<unsigned char>( *first ) );
+        ++first;
+
+        if( first == last )
+        {
+            hash_combine( seed, w | 0x0100u );
+            return;
+        }
+
+        w |= static_cast<boost::uint32_t>( static_cast<unsigned char>( *first ) ) << 8;
+        ++first;
+
+        if( first == last )
+        {
+            hash_combine( seed, w | 0x010000u );
+            return;
+        }
+
+        w |= static_cast<boost::uint32_t>( static_cast<unsigned char>( *first ) ) << 16;
+        ++first;
+
+        if( first == last )
+        {
+            hash_combine( seed, w | 0x01000000u );
+            return;
+        }
+
+        w |= static_cast<boost::uint32_t>( static_cast<unsigned char>( *first ) ) << 24;
+        ++first;
+
+        hash_combine( seed, w );
+    }
+}
+
+} // namespace hash_detail
+} // namespace boost
+
+#endif // #ifndef BOOST_HASH_DETAIL_HASH_RANGE_HPP
diff --git a/include/boost/container_hash/hash.hpp b/include/boost/container_hash/hash.hpp
index 090250b..56bfa01 100644
--- a/include/boost/container_hash/hash.hpp
+++ b/include/boost/container_hash/hash.hpp
@@ -16,6 +16,7 @@
 #include <boost/container_hash/is_unordered_range.hpp>
 #include <boost/container_hash/detail/hash_tuple.hpp>
 #include <boost/container_hash/detail/hash_mix.hpp>
+#include <boost/container_hash/detail/hash_range.hpp>
 #include <boost/type_traits/is_enum.hpp>
 #include <boost/type_traits/is_integral.hpp>
 #include <boost/type_traits/is_floating_point.hpp>
@@ -559,10 +560,7 @@ namespace boost
     template <class It>
     inline void hash_range( std::size_t& seed, It first, It last )
     {
-        for( ; first != last; ++first )
-        {
-            hash_combine<typename std::iterator_traits<It>::value_type>( seed, *first );
-        }
+        hash_detail::hash_range( seed, first, last );
     }
 
     template <class It>