[GR-32616] [GR-32512] GC improvements and adaptive policy prototype.

PullRequest: graal/9358

Author: peter-hofer

substratevm/src/com.oracle.svm.core.genscavenge/src/com/oracle/svm/core/genscavenge/AbstractCollectionPolicy.java

@@ -0,0 +1,138 @@
+/*
+ * Copyright (c) 2021, 2021, Oracle and/or its affiliates. All rights reserved.
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This code is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 only, as
+ * published by the Free Software Foundation.  Oracle designates this
+ * particular file as subject to the "Classpath" exception as provided
+ * by Oracle in the LICENSE file that accompanied this code.
+ *
+ * This code is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ * version 2 for more details (a copy is included in the LICENSE file that
+ * accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License version
+ * 2 along with this work; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
+ * or visit www.oracle.com if you need additional information or have any
+ * questions.
+ */
+package com.oracle.svm.core.genscavenge;
+
+import org.graalvm.word.UnsignedWord;
+
+import com.oracle.svm.core.util.UnsignedUtils;
+
+/**
+ * A weighted average maintains a running, weighted average of some floating-point value.
+ *
+ * The average is adaptive in that we smooth it for the initial samples; we don't use the weight
+ * until we have enough samples for it to be meaningful.
+ *
+ * This serves as our best estimate of a future unknown.
+ */
+class AdaptiveWeightedAverage {
+    static final int OLD_THRESHOLD = 100;
+
+    private final int weight;
+
+    private double average;
+    private long sampleCount;
+    private boolean isOld;
+
+    AdaptiveWeightedAverage(int weight) {
+        this(weight, 0);
+    }
+
+    AdaptiveWeightedAverage(int weight, double avg) {
+        this.weight = weight;
+        this.average = avg;
+    }
+
+    public double getAverage() {
+        return average;
+    }
+
+    public void sample(double value) {
+        sampleCount++;
+        if (!isOld && sampleCount > OLD_THRESHOLD) {
+            isOld = true;
+        }
+        average = computeAdaptiveAverage(value, average);
+    }
+
+    public final void sample(UnsignedWord value) {
+        sample(UnsignedUtils.toDouble(value));
+    }
+
+    protected double computeAdaptiveAverage(double sample, double avg) {
+        /*
+         * We smoothen the samples by not using weight directly until we've had enough data to make
+         * it meaningful. We'd like the first weight used to be 1, the second to be 1/2, etc until
+         * we have OLD_THRESHOLD/weight samples.
+         */
+        long countWeight = 0;
+        if (!isOld) { // avoid division by zero if the counter wraps
+            countWeight = OLD_THRESHOLD / sampleCount;
+        }
+        long adaptiveWeight = Math.max(weight, countWeight);
+        return expAvg(avg, sample, adaptiveWeight);
+    }
+
+    private static double expAvg(double avg, double sample, long adaptiveWeight) {
+        assert adaptiveWeight <= 100 : "weight must be a percentage";
+        return (100.0 - adaptiveWeight) * avg / 100.0 + adaptiveWeight * sample / 100.0;
+    }
+}
+
+/**
+ * A weighted average that includes a deviation from the average, some multiple of which is added to
+ * the average.
+ *
+ * This serves as our best estimate of an upper bound on a future unknown.
+ */
+class AdaptivePaddedAverage extends AdaptiveWeightedAverage {
+    private final int padding;
+    private final boolean noZeroDeviations;
+
+    private double paddedAverage;
+    private double deviation;
+
+    AdaptivePaddedAverage(int weight, int padding) {
+        this(weight, padding, false);
+    }
+
+    /**
+     * @param noZeroDeviations do not update deviations when a sample is zero. The average is
+     *            allowed to change. This is to prevent zero samples from drastically changing the
+     *            padded average.
+     */
+    AdaptivePaddedAverage(int weight, int padding, boolean noZeroDeviations) {
+        super(weight);
+        this.padding = padding;
+        this.noZeroDeviations = noZeroDeviations;
+    }
+
+    @Override
+    public void sample(double value) {
+        super.sample(value);
+        double average = super.getAverage();
+        if (value != 0 || !noZeroDeviations) {
+            deviation = computeAdaptiveAverage(Math.abs(value - average), deviation);
+        }
+        paddedAverage = average + padding * deviation;
+    }
+
+    public double getPaddedAverage() {
+        return paddedAverage;
+    }
+
+    public double getDeviation() {
+        return deviation;
+    }
+}

substratevm/src/com.oracle.svm.core.genscavenge/src/com/oracle/svm/core/genscavenge/AdaptiveCollectionPolicy.java

@@ -124,7 +124,7 @@ public static AlignedHeader getEnclosingChunk(Object obj) {
     }
 
     public static AlignedHeader getEnclosingChunkFromObjectPointer(Pointer ptr) {
-        return (AlignedHeader) PointerUtils.roundDown(ptr, HeapPolicy.getAlignedHeapChunkAlignment());
+        return (AlignedHeader) PointerUtils.roundDown(ptr, HeapParameters.getAlignedHeapChunkAlignment());
     }
 
     /** Return the offset of an object within the objects part of a chunk. */