[GR-40382] Add storage strategy when creating list or tuple from bytecode stack

PullRequest: graalpython/2398

Author: tomasstupka

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/list/PList.java

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017, 2021, Oracle and/or its affiliates.
+ * Copyright (c) 2017, 2022, Oracle and/or its affiliates.
  * Copyright (c) 2013, Regents of the University of California
  *
  * All rights reserved.
@@ -30,10 +30,10 @@
 import com.oracle.graal.python.builtins.objects.ints.PInt;
 import com.oracle.graal.python.nodes.ErrorMessages;
 import com.oracle.graal.python.nodes.interop.PForeignToPTypeNode;
-import com.oracle.graal.python.nodes.literal.ListLiteralNode;
 import com.oracle.graal.python.runtime.GilNode;
 import com.oracle.graal.python.runtime.exception.PException;
 import com.oracle.graal.python.runtime.sequence.PSequence;
+import com.oracle.graal.python.runtime.sequence.storage.BasicSequenceStorage;
 import com.oracle.graal.python.runtime.sequence.storage.SequenceStorage;
 import com.oracle.graal.python.util.OverflowException;
 import com.oracle.truffle.api.CompilerAsserts;
@@ -51,7 +51,7 @@
 
 @ExportLibrary(InteropLibrary.class)
 public final class PList extends PSequence {
-    private final ListLiteralNode origin;
+    private final ListOrigin origin;
     private SequenceStorage store;
 
     public PList(Object cls, Shape instanceShape, SequenceStorage store) {
@@ -60,7 +60,7 @@ public PList(Object cls, Shape instanceShape, SequenceStorage store) {
         this.store = store;
     }
 
-    public PList(Object cls, Shape instanceShape, SequenceStorage store, ListLiteralNode origin) {
+    public PList(Object cls, Shape instanceShape, SequenceStorage store, ListOrigin origin) {
         super(cls, instanceShape);
         this.origin = origin;
         this.store = store;
@@ -117,15 +117,15 @@ public final int hashCode() {
         return super.hashCode();
     }
 
-    public ListLiteralNode getOrigin() {
+    public ListOrigin getOrigin() {
         return origin;
     }
 
     @ExportMessage
     public SourceSection getSourceLocation(@Exclusive @Cached GilNode gil) throws UnsupportedMessageException {
         boolean mustRelease = gil.acquire();
         try {
-            ListLiteralNode node = getOrigin();
+            ListOrigin node = getOrigin();
             SourceSection result = null;
             if (node != null) {
                 result = node.getSourceSection();
@@ -229,4 +229,38 @@ public void removeArrayElement(long index,
             gil.release(mustRelease);
         }
     }
+
+    public interface ListOrigin {
+
+        /**
+         * This class serves the purpose of updating the size estimate for the lists constructed
+         * here over time. The estimate is updated slowly, it takes {@link #NUM_DIGITS_POW2} lists
+         * to reach a size one larger than the current estimate to increase the estimate for new
+         * lists.
+         */
+        @CompilerDirectives.ValueType
+        public static final class SizeEstimate {
+            private static final int NUM_DIGITS = 3;
+            private static final int NUM_DIGITS_POW2 = 1 << NUM_DIGITS;
+
+            @CompilerDirectives.CompilationFinal private int shiftedStorageSizeEstimate;
+
+            public SizeEstimate(int storageSizeEstimate) {
+                shiftedStorageSizeEstimate = storageSizeEstimate * NUM_DIGITS_POW2;
+            }
+
+            public int estimate() {
+                return shiftedStorageSizeEstimate >> NUM_DIGITS;
+            }
+
+            public int updateFrom(int newSizeEstimate) {
+                shiftedStorageSizeEstimate = shiftedStorageSizeEstimate + newSizeEstimate - estimate();
+                return shiftedStorageSizeEstimate;
+            }
+        }
+
+        void reportUpdatedCapacity(BasicSequenceStorage newStore);
+
+        SourceSection getSourceSection();
+    }
 }

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/nodes/builtins/ListNodes.java

@@ -40,10 +40,10 @@
  */
 package com.oracle.graal.python.nodes.builtins;
 
-import static com.oracle.graal.python.runtime.exception.PythonErrorType.TypeError;
+import static com.oracle.graal.python.builtins.objects.str.StringUtils.cat;
 import static com.oracle.graal.python.nodes.BuiltinNames.T_LIST;
 import static com.oracle.graal.python.nodes.StringLiterals.T_SPACE;
-import static com.oracle.graal.python.builtins.objects.str.StringUtils.cat;
+import static com.oracle.graal.python.runtime.exception.PythonErrorType.TypeError;
 
 import com.oracle.graal.python.builtins.PythonBuiltinClassType;
 import com.oracle.graal.python.builtins.modules.MathGuards;
@@ -67,7 +67,6 @@
 import com.oracle.graal.python.nodes.builtins.ListNodesFactory.FastConstructListNodeGen;
 import com.oracle.graal.python.nodes.builtins.ListNodesFactory.IndexNodeGen;
 import com.oracle.graal.python.nodes.call.special.LookupAndCallUnaryNode;
-import com.oracle.graal.python.nodes.literal.ListLiteralNode;
 import com.oracle.graal.python.nodes.object.GetClassNode;
 import com.oracle.graal.python.nodes.truffle.PythonArithmeticTypes;
 import com.oracle.graal.python.nodes.util.CastToTruffleStringNode;
@@ -289,7 +288,7 @@ Object doGeneric(VirtualFrame frame, Object object) {
      * interpreter, it will use a different {@link AppendNode} than in the compiled code, so the
      * specializations for the compiler are different. This is useful, because in the interpreter we
      * will report back type and size changes to the origin of the list via {@link PList#getOrigin}
-     * and {@link ListLiteralNode#reportUpdatedCapacity}. Thus, there's a chance that the compiled
+     * and {@link PList.ListOrigin#reportUpdatedCapacity}. Thus, there's a chance that the compiled
      * code will only see lists of the correct size and storage type.
      */
     @GenerateUncached

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/nodes/bytecode/PBytecodeRootNode.java

@@ -127,6 +127,10 @@
 import com.oracle.graal.python.nodes.builtins.ListNodesFactory;
 import com.oracle.graal.python.nodes.builtins.TupleNodes;
 import com.oracle.graal.python.nodes.builtins.TupleNodesFactory;
+import com.oracle.graal.python.nodes.bytecode.SequenceFromStackNode.ListFromStackNode;
+import com.oracle.graal.python.nodes.bytecode.SequenceFromStackNode.TupleFromStackNode;
+import com.oracle.graal.python.nodes.bytecode.SequenceFromStackNodeFactory.ListFromStackNodeGen;
+import com.oracle.graal.python.nodes.bytecode.SequenceFromStackNodeFactory.TupleFromStackNodeGen;
 import com.oracle.graal.python.nodes.call.CallNode;
 import com.oracle.graal.python.nodes.call.CallNodeGen;
 import com.oracle.graal.python.nodes.call.special.CallBinaryMethodNode;
@@ -4240,14 +4244,14 @@ private int bytecodeCollectionFromStack(VirtualFrame virtualFrame, int type, int
         Object res = null;
         switch (type) {
             case CollectionBits.KIND_LIST: {
-                Object[] store = new Object[count];
-                moveFromStack(virtualFrame, stackTop - count + 1, stackTop + 1, store);
-                res = factory.createList(store);
+                ListFromStackNode storageFromStackNode = insertChildNodeInt(localNodes, nodeIndex, ListFromStackNodeGen.class, (int length) -> ListFromStackNodeGen.create(length), count);
+                SequenceStorage store = storageFromStackNode.execute(virtualFrame, stackTop - count + 1, stackTop + 1);
+                res = factory.createList(store, storageFromStackNode);
                 break;
             }
             case CollectionBits.KIND_TUPLE: {
-                Object[] store = new Object[count];
-                moveFromStack(virtualFrame, stackTop - count + 1, stackTop + 1, store);
+                TupleFromStackNode storageFromStackNode = insertChildNodeInt(localNodes, nodeIndex, TupleFromStackNodeGen.class, (int length) -> TupleFromStackNodeGen.create(length), count);
+                SequenceStorage store = storageFromStackNode.execute(virtualFrame, stackTop - count + 1, stackTop + 1);
                 res = factory.createTuple(store);
                 break;
             }