Shave 39% from snapshot size.

src/serialize.cc

Index: src/serialize.cc
===================================================================
--- src/serialize.cc	(revision 9713)
+++ src/serialize.cc	(working copy)
@@ -883,12 +883,18 @@
   CASE_STATEMENT(where, how, within, kLargeCode)                               \
   CASE_BODY(where, how, within, kLargeCode, kUnknownOffsetFromStart)
 
-#define EMIT_COMMON_REFERENCE_PATTERNS(pseudo_space_number,                    \
-                                       space_number,                           \
-                                       offset_from_start)                      \
-  CASE_STATEMENT(kFromStart, kPlain, kStartOfObject, pseudo_space_number)      \
-  CASE_BODY(kFromStart, kPlain, kStartOfObject, space_number, offset_from_start)
+#define FOUR_CASES(byte_code)             \
+  case byte_code:                         \
+  case byte_code + 1:                     \
+  case byte_code + 2:                     \
+  case byte_code + 3:
 
+#define SIXTEEN_CASES(byte_code)          \
+  FOUR_CASES(byte_code)                   \
+  FOUR_CASES(byte_code + 4)               \
+  FOUR_CASES(byte_code + 8)               \
+  FOUR_CASES(byte_code + 12)
+
       // We generate 15 cases and bodies that process special tags that combine
       // the raw data tag and the length into one byte.
 #define RAW_CASE(index, size)                                      \
@@ -911,6 +917,32 @@
         break;
       }
 
+      SIXTEEN_CASES(kRootArrayLowConstants)
+      SIXTEEN_CASES(kRootArrayHighConstants) {
+        int root_id = RootArrayConstantFromByteCode(data);
+        *current++ = isolate->heap()->roots_address()[root_id];

[Lasse Reichstein, 2011/10/20 10:06:54]

The name roots_address suggests that it returns an Address. Could it be called
roots_array_start or _origin instead?

[Erik Corry, 2011/10/20 11:01:42]

Done.

+        break;
+      }
+
+      case kRepeat: {
+        int repeats = source_->GetInt();
+        Object* object = current[-1];
+        for (int i = 0; i < repeats; i++) current[i] = object;
+        current += repeats;
+        break;
+      }
+
+      STATIC_ASSERT(kMaxRepeats == 12);
+      FOUR_CASES(kConstantRepeat)
+      FOUR_CASES(kConstantRepeat + 4)
+      FOUR_CASES(kConstantRepeat + 8) {
+        int repeats = RepeatsForCode(data);
+        Object* object = current[-1];
+        for (int i = 0; i < repeats; i++) current[i] = object;
+        current += repeats;
+        break;
+      }
+
       // Deserialize a new object and write a pointer to it to the current
       // object.
       ONE_PER_SPACE(kNewObject, kPlain, kStartOfObject)
@@ -936,9 +968,6 @@
       // start and write a pointer to its first instruction to the current code
       // object.
       ALL_SPACES(kFromStart, kFromCode, kFirstInstruction)
-      // Find an already deserialized object at one of the predetermined popular
-      // offsets from the start and write a pointer to it in the current object.
-      COMMON_REFERENCE_PATTERNS(EMIT_COMMON_REFERENCE_PATTERNS)
       // Find an object in the roots array and write a pointer to it to the
       // current object.
       CASE_STATEMENT(kRootArray, kPlain, kStartOfObject, 0)
@@ -980,7 +1009,6 @@
 #undef CASE_BODY
 #undef ONE_PER_SPACE
 #undef ALL_SPACES
-#undef EMIT_COMMON_REFERENCE_PATTERNS
 #undef ASSIGN_DEST_SPACE
 
       case kNewPage: {
@@ -1067,7 +1095,8 @@
     : sink_(sink),
       current_root_index_(0),
       external_reference_encoder_(new ExternalReferenceEncoder),
-      large_object_total_(0) {
+      large_object_total_(0),
+      root_index_wave_front_(0) {
   // The serializer is meant to be used only to generate initial heap images
   // from a context in which there is only one isolate.
   ASSERT(Isolate::Current()->IsDefaultIsolate());
@@ -1124,6 +1153,9 @@
   Isolate* isolate = Isolate::Current();
 
   for (Object** current = start; current < end; current++) {
+    if (start == isolate->heap()->roots_address()) {
+      root_index_wave_front_ = Max(root_index_wave_front_, current - start);
+    }
     if (reinterpret_cast<Address>(current) ==
         isolate->heap()->store_buffer()->TopAddress()) {
       sink_->Put(kSkip, "Skip");
@@ -1191,10 +1223,12 @@
 }
 
 
-int PartialSerializer::RootIndex(HeapObject* heap_object) {
-  for (int i = 0; i < Heap::kRootListLength; i++) {
-    Object* root = HEAP->roots_address()[i];
-    if (root == heap_object) return i;
+int Serializer::RootIndex(HeapObject* heap_object) {
+  Heap* heap = HEAP;
+  if (heap->InNewSpace(heap_object)) return kInvalidRootIndex;
+  for (int i = 0; i < root_index_wave_front_; i++) {
+    Object* root = heap->roots_address()[i];
+    if (!root->IsSmi() && root == heap_object) return i;
   }
   return kInvalidRootIndex;
 }
@@ -1230,18 +1264,8 @@
   // all objects) then we should shift out the bits that are always 0.
   if (!SpaceIsLarge(space)) address >>= kObjectAlignmentBits;
   if (from_start) {
-#define COMMON_REFS_CASE(pseudo_space, actual_space, offset)                   \
-    if (space == actual_space && address == offset &&                          \
-        how_to_code == kPlain && where_to_point == kStartOfObject) {           \
-      sink_->Put(kFromStart + how_to_code + where_to_point +                   \
-                 pseudo_space, "RefSer");                                      \
-    } else  /* NOLINT */
-    COMMON_REFERENCE_PATTERNS(COMMON_REFS_CASE)
-#undef COMMON_REFS_CASE
-    {  /* NOLINT */
-      sink_->Put(kFromStart + how_to_code + where_to_point + space, "RefSer");
-      sink_->PutInt(address, "address");
-    }
+    sink_->Put(kFromStart + how_to_code + where_to_point + space, "RefSer");
+    sink_->PutInt(address, "address");
   } else {
     sink_->Put(kBackref + how_to_code + where_to_point + space, "BackRefSer");
     sink_->PutInt(address, "address");
@@ -1256,6 +1280,12 @@
   CHECK(o->IsHeapObject());
   HeapObject* heap_object = HeapObject::cast(o);
 
+  int root_index;
+  if ((root_index = RootIndex(heap_object)) != kInvalidRootIndex) {
+    PutRoot(root_index, heap_object, how_to_code, where_to_point);
+    return;
+  }
+
   if (address_mapper_.IsMapped(heap_object)) {
     int space = SpaceOfAlreadySerializedObject(heap_object);
     int address = address_mapper_.MappedTo(heap_object);
@@ -1286,6 +1316,26 @@
 }
 
 
+void Serializer::PutRoot(int root_index,
+                         HeapObject* object,
+                         SerializerDeserializer::HowToCode how_to_code,
+                         SerializerDeserializer::WhereToPoint where_to_point) {
+  if (how_to_code == kPlain &&
+      where_to_point == kStartOfObject &&
+      root_index < 0x20 &&

[Lasse Reichstein, 2011/10/20 10:06:54]

Name the constants.

[Erik Corry, 2011/10/20 11:01:42]

Done.

+      !HEAP->InNewSpace(object)) {
+    if (root_index < 0x10) {
+      sink_->Put(kRootArrayLowConstants + root_index, "RootLoConstant");
+    } else {
+      sink_->Put(kRootArrayHighConstants + root_index - 0x10, "RootHiConstant");
+    }
+  } else {
+    sink_->Put(kRootArray + how_to_code + where_to_point, "RootSerialization");
+    sink_->PutInt(root_index, "root_index");
+  }
+}
+
+
 void PartialSerializer::SerializeObject(
     Object* o,
     HowToCode how_to_code,
@@ -1295,8 +1345,7 @@
 
   int root_index;
   if ((root_index = RootIndex(heap_object)) != kInvalidRootIndex) {
-    sink_->Put(kRootArray + how_to_code + where_to_point, "RootSerialization");
-    sink_->PutInt(root_index, "root_index");
+    PutRoot(root_index, heap_object, how_to_code, where_to_point);
     return;
   }
 
@@ -1374,9 +1423,26 @@
     if (current < end) OutputRawData(reinterpret_cast<Address>(current));
 
     while (current < end && !(*current)->IsSmi()) {
-      serializer_->SerializeObject(*current, kPlain, kStartOfObject);
-      bytes_processed_so_far_ += kPointerSize;
-      current++;
+      if (current != start &&
+          current[0] == current[-1] &&
+          !HEAP->InNewSpace(*current)) {
+        int repeat_count = 1;
+        while (current < end - 1 && current[repeat_count] == current[0]) {
+          repeat_count++;
+        }
+        current += repeat_count;
+        bytes_processed_so_far_ += repeat_count * kPointerSize;
+        if (repeat_count > kMaxRepeats) {

[Lasse Reichstein, 2011/10/20 10:06:54]

A generic repeat takes five bytes. Up to 4*kMaxRepeats, it's smaller to do
repeated constant repeats.

[Erik Corry, 2011/10/20 11:01:42]

No, a generic repeat takes 2 bytes if the repeat count is less than 128.

+          sink_->Put(kRepeat, "SerializeRepeats");
+          sink_->PutInt(repeat_count, "SerializeRepeats");
+        } else {
+          sink_->Put(CodeForRepeats(repeat_count), "SerializeRepeats");
+        }
+      } else {
+        serializer_->SerializeObject(*current, kPlain, kStartOfObject);
+        bytes_processed_so_far_ += kPointerSize;
+        current++;
+      }
     }
   }
 }