Implement fast path in UTF8ToUTF16 for pure ASCII strings

base/strings/string_util.cc

Index: base/strings/string_util.cc
diff --git a/base/strings/string_util.cc b/base/strings/string_util.cc
index ab9570f6741a5f8d981dbfd278fa37512320da50..87d51336ffacbf604823c9ab7a2dd08614bf4bf9 100644
--- a/base/strings/string_util.cc
+++ b/base/strings/string_util.cc
@@ -64,6 +64,34 @@ static bool CompareParameter(const ReplacementOffset& elem1,
   return elem1.parameter < elem2.parameter;
 }
 
+// Assuming that a pointer is the size of a "machine word", then
+// uintptr_t is an integer type that is also a machine word.
+typedef uintptr_t MachineWord;
+const uintptr_t machineWordAlignmentMask = sizeof(MachineWord) - 1;

[brettw, 2014/09/16 17:14:28]

Constant naming: kMachineWord....

[mnaganov (inactive), 2014/09/17 15:15:01]

Done.

+
+inline bool IsAlignedToMachineWord(const void* pointer) {
+    return !(reinterpret_cast<MachineWord>(pointer) & machineWordAlignmentMask);
+}
+
+template<typename T> inline T* AlignToMachineWord(T* pointer) {
+  return reinterpret_cast<T*>(reinterpret_cast<MachineWord>(pointer) &
+                              ~machineWordAlignmentMask);
+}
+
+template<size_t size, typename CharacterType> struct NonASCIIMask;
+template<> struct NonASCIIMask<4, base::char16> {
+    static inline uint32_t value() { return 0xFF80FF80U; }
+};
+template<> struct NonASCIIMask<4, char> {
+    static inline uint32_t value() { return 0x80808080U; }
+};
+template<> struct NonASCIIMask<8, base::char16> {
+    static inline uint64_t value() { return 0xFF80FF80FF80FF80ULL; }
+};
+template<> struct NonASCIIMask<8, char> {
+    static inline uint64_t value() { return 0x8080808080808080ULL; }
+};
+
 }  // namespace
 
 namespace base {
@@ -322,22 +350,50 @@ bool ContainsOnlyChars(const StringPiece16& input,
   return input.find_first_not_of(characters) == StringPiece16::npos;
 }
 
-template<class STR>
-static bool DoIsStringASCII(const STR& str) {
-  for (size_t i = 0; i < str.length(); i++) {
-    typename ToUnsigned<typename STR::value_type>::Unsigned c = str[i];
-    if (c > 0x7F)
-      return false;
+template <class Char>
+inline bool DoIsStringASCII(const Char* characters, size_t length) {
+  MachineWord all_char_bits = 0;
+  const Char* end = characters + length;
+
+  // Prologue: align the input.
+  while (!IsAlignedToMachineWord(characters) && characters != end) {
+    all_char_bits |= *characters;
+    ++characters;
   }
-  return true;
+
+  // Compare the values of CPU word size.
+  const Char* word_end = AlignToMachineWord(end);
+  const size_t loop_increment = sizeof(MachineWord) / sizeof(Char);
+  while (characters < word_end) {
+    all_char_bits |= *(reinterpret_cast<const MachineWord*>(characters));
+    characters += loop_increment;
+  }
+
+  // Process the remaining bytes.
+  while (characters != end) {
+    all_char_bits |= *characters;
+    ++characters;
+  }
+
+  MachineWord non_ascii_bit_mask =
+      NonASCIIMask<sizeof(MachineWord), Char>::value();
+  return !(all_char_bits & non_ascii_bit_mask);
 }
 
 bool IsStringASCII(const StringPiece& str) {
-  return DoIsStringASCII(str);
+  return DoIsStringASCII(str.data(), str.length());
 }
 
 bool IsStringASCII(const string16& str) {
-  return DoIsStringASCII(str);
+  return DoIsStringASCII(str.data(), str.length());
+}
+
+bool IsStringASCII(const char* src, size_t src_len) {
+  return DoIsStringASCII(src, src_len);
+}
+
+bool IsStringASCII(const char16* src, size_t src_len) {
+  return DoIsStringASCII(src, src_len);
 }
 
 bool IsStringUTF8(const std::string& str) {