Support simulations of large caches in weighted policy types

In weight-based traces the entries have different sizes, e.g. bytes. The
policies that support this characteristic will evict multiple entries
until at its maximum-size contraint. This was an integer setting which
is too small for the Tencent Photo trace.

The setting was increased to a long and the weighted traces updated
accordingly. The other traces use a checked cast to an int value as a
minimal migration to minimize the chance of introducing bugs. The
linked policy was fixed to handle the entry's weight changing on a
cache hit.

The AdaptSize simulator compatibility was reduced to only its trace
reader. That simulator treats the same key with different weights as
distinct entries, rather than an update. The workaround of hashing
the key and weight as the entry's key causes noticable workload skew.
Since other simulators (like LHD from the same group) don't make this
odd choice, its now restricted to just their trace reader.

gradle/dependencies.gradle

@@ -26,7 +26,7 @@
 ext {
   versions = [
     akka: '2.6.8',
-    cache2k: '1.3.2.Alpha',
+    cache2k: '1.3.3.Alpha',
     checkerFramework: '3.6.0',
     coherence: '20.06',
     collision: '0.3.3',

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/BasicSettings.java

@@ -76,8 +76,8 @@ public TinyLfuSettings tinyLfu() {
     return new TinyLfuSettings();
   }
 
-  public int maximumSize() {
-    return config().getInt("maximum-size");
+  public long maximumSize() {
+    return config().getLong("maximum-size");
   }
 
   public boolean isFiles() {

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/admission/countmin64/CountMin64TinyLfu.java

@@ -17,6 +17,7 @@
 
 import com.github.benmanes.caffeine.cache.simulator.BasicSettings;
 import com.github.benmanes.caffeine.cache.simulator.admission.Frequency;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 /**
@@ -45,8 +46,8 @@ public CountMin64TinyLfu(Config config) {
     BasicSettings settings = new BasicSettings(config);
     sketch = new CountMin64(settings.tinyLfu().countMin64().eps(),
         settings.tinyLfu().countMin64().confidence(), settings.randomSeed());
+    sampleSize = Ints.checkedCast(10 * settings.maximumSize());
     conservative = settings.tinyLfu().conservative();
-    sampleSize = 10 * settings.maximumSize();
   }
 
   /** Returns the estimated usage frequency of the item. */

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/admission/perfect/PerfectFrequency.java

@@ -17,6 +17,7 @@
 
 import com.github.benmanes.caffeine.cache.simulator.BasicSettings;
 import com.github.benmanes.caffeine.cache.simulator.admission.Frequency;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2IntMap;
@@ -34,7 +35,8 @@ public final class PerfectFrequency implements Frequency {
   private int size;
 
   public PerfectFrequency(Config config) {
-    sampleSize = 10 * new BasicSettings(config).maximumSize();
+    BasicSettings settings = new BasicSettings(config);
+    sampleSize = Ints.checkedCast(10 * settings.maximumSize());
     counts = new Long2IntOpenHashMap();
   }
 

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/admission/table/RandomRemovalFrequencyTable.java

@@ -23,6 +23,7 @@
 
 import com.github.benmanes.caffeine.cache.simulator.BasicSettings;
 import com.github.benmanes.caffeine.cache.simulator.admission.Frequency;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 /**
@@ -49,7 +50,7 @@ public final class RandomRemovalFrequencyTable implements Frequency {
 
   public RandomRemovalFrequencyTable(Config config) {
     BasicSettings settings = new BasicSettings(config);
-    maxSum = sampleFactor * settings.maximumSize();
+    maxSum = Ints.checkedCast(sampleFactor * settings.maximumSize());
     random = new Random(settings.randomSeed());
     table = new HashMap<>(maxSum);
   }

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/admission/tinycache/TinyCacheAdapter.java

@@ -17,6 +17,7 @@
 
 import com.github.benmanes.caffeine.cache.simulator.BasicSettings;
 import com.github.benmanes.caffeine.cache.simulator.admission.Frequency;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 /**
@@ -37,7 +38,8 @@ public final class TinyCacheAdapter implements Frequency {
    */
   public TinyCacheAdapter(Config config) {
     BasicSettings settings = new BasicSettings(config);
-    int nrSets = sampleFactor * settings.maximumSize() / 64; // number of (independent sets)
+    // number of (independent sets)
+    int nrSets = Ints.checkedCast(sampleFactor * settings.maximumSize() / 64);
     tcs = new TinyCacheSketch(nrSets, 64,settings.randomSeed());
   }
 

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/parser/adapt_size/AdaptSizeTraceReader.java

@@ -25,10 +25,14 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.AccessEvent;
 import com.github.benmanes.caffeine.cache.simulator.policy.Policy.Characteristic;
 import com.google.common.collect.Sets;
+import com.google.common.hash.Hashing;
 
 /**
  * A reader for the trace files provided by the authors of the AdaptSize algorithm. See
  * <a href="https://github.com/dasebe/webcachesim#how-to-get-traces">traces</a>.
+ * <p>
+ * The AdaptSize simulator treats identical keys with different weights as unique entries, rather
+ * than as an update to that entry's size. This behavior is emulated by a key hash.
  *
  * @author [email protected] (Ben Manes)
  */
@@ -47,7 +51,12 @@ public Set<Characteristic> characteristics() {
   public Stream<AccessEvent> events() throws IOException {
     return lines()
         .map(line -> line.split(" ", 3))
-        .map(array -> AccessEvent.forKeyAndWeight(
-            Long.parseLong(array[1]), Integer.parseInt(array[2])));
+        .map(array -> {
+          long key = Long.parseLong(array[1]);
+          int weight = Integer.parseInt(array[2]);
+          long hashKey = Hashing.murmur3_128().newHasher()
+              .putLong(key).putInt(weight).hash().asLong();
+          return AccessEvent.forKeyAndWeight(hashKey, weight);
+        });
   }
 }

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/AccessEvent.java

@@ -101,7 +101,7 @@ private static final class WeightedAccessEvent extends AccessEvent {
     private final int weight;
 
     WeightedAccessEvent(long key, int weight) {
-      super(cantorHashCode(key, weight));
+      super(key);
       this.weight = weight;
       checkArgument(weight >= 0);
     }
@@ -112,11 +112,6 @@ public int weight() {
     }
   }
 
-  /** Cantor pairing function. */
-  private static long cantorHashCode(long key, int weight) {
-    return (key + weight) * (key + weight + 1) / 2 + weight;
-  }
-
   private static final class PenaltiesAccessEvent extends AccessEvent {
     private final double missPenalty;
     private final double hitPenalty;

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/adaptive/ArcPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -75,8 +76,8 @@ public final class ArcPolicy implements KeyOnlyPolicy {
 
   public ArcPolicy(Config config) {
     BasicSettings settings = new BasicSettings(config);
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
     this.policyStats = new PolicyStats("adaptive.Arc");
-    this.maximumSize = settings.maximumSize();
     this.data = new Long2ObjectOpenHashMap<>();
     this.headT1 = new Node();
     this.headT2 = new Node();

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/adaptive/CarPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -61,9 +62,9 @@ public final class CarPolicy implements KeyOnlyPolicy {
 
   public CarPolicy(Config config) {
     BasicSettings settings = new BasicSettings(config);
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
     this.policyStats = new PolicyStats("adaptive.Car");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headT1 = new Node();
     this.headT2 = new Node();
     this.headB1 = new Node();

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/adaptive/CartPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -65,9 +66,9 @@ public final class CartPolicy implements KeyOnlyPolicy {
 
   public CartPolicy(Config config) {
     BasicSettings settings = new BasicSettings(config);
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
     this.policyStats = new PolicyStats("adaptive.Cart");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headT1 = new Node();
     this.headT2 = new Node();
     this.headB1 = new Node();

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/ClockProPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.Policy.KeyOnlyPolicy;
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -86,10 +87,10 @@ public final class ClockProPolicy implements KeyOnlyPolicy {
 
   public ClockProPolicy(Config config) {
     BasicSettings settings = new BasicSettings(config);
+    maximumSize = Ints.checkedCast(settings.maximumSize());
     policyStats = new PolicyStats("irr.ClockPro");
-    maximumColdSize = settings.maximumSize();
     data = new Long2ObjectOpenHashMap<>();
-    maximumSize = settings.maximumSize();
+    maximumColdSize = maximumSize;
 
     // All the hands move in the clockwise direction
     handHot = handCold = handTest = null;

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/DClockPolicy.java

@@ -27,6 +27,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.Policy.KeyOnlyPolicy;
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -66,9 +67,9 @@ public final class DClockPolicy implements KeyOnlyPolicy {
 
   public DClockPolicy(DClockSettings settings, double percentActive) {
     this.policyStats = new PolicyStats("irr.DClock (active: %d%%)", (int) (100 * percentActive));
-    this.maxActive = (int) (percentActive * settings.maximumSize());
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
+    this.maxActive = (int) (percentActive * maximumSize);
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headNonResident = new Node();
     this.headInactive = new Node();
     this.headActive = new Node();

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/FrdPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectOpenHashMap;
@@ -62,11 +63,11 @@ public final class FrdPolicy implements KeyOnlyPolicy {
 
   public FrdPolicy(Config config) {
     FrdSettings settings = new FrdSettings(config);
-    this.maximumMainResidentSize = (int) (settings.maximumSize() * settings.percentMain());
-    this.maximumFilterSize = settings.maximumSize() - maximumMainResidentSize;
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
+    this.maximumMainResidentSize = (int) (maximumSize * settings.percentMain());
+    this.maximumFilterSize = maximumSize - maximumMainResidentSize;
     this.policyStats = new PolicyStats("irr.Frd");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headFilter = new Node();
     this.headMain = new Node();
   }

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/HillClimberFrdPolicy.java

@@ -25,6 +25,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectOpenHashMap;
@@ -60,16 +61,16 @@ public final class HillClimberFrdPolicy implements KeyOnlyPolicy {
 
   public HillClimberFrdPolicy(Config config) {
     FrdSettings settings = new FrdSettings(config);
-    this.maximumMainResidentSize = (int) (settings.maximumSize() * settings.percentMain());
-    this.maximumFilterSize = settings.maximumSize() - maximumMainResidentSize;
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
+    this.maximumMainResidentSize = (int) (maximumSize * settings.percentMain());
+    this.maximumFilterSize = maximumSize - maximumMainResidentSize;
     this.policyStats = new PolicyStats("irr.AdaptiveFrd");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headFilter = new Node();
     this.headMain = new Node();
 
-    this.sampleSize = 10 * settings.maximumSize();
-    this.pivot = (int) (0.05 * settings.maximumSize());
+    this.pivot = (int) (0.05 * maximumSize);
+    this.sampleSize = 10 * maximumSize;
     this.tolerance = 100d * 0;
   }
 

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/IndicatorFrdPolicy.java

@@ -26,6 +26,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.sketch.Indicator;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectOpenHashMap;
@@ -54,11 +55,11 @@ public final class IndicatorFrdPolicy implements KeyOnlyPolicy {
   public IndicatorFrdPolicy(Config config) {
     FrdSettings settings = new FrdSettings(config);
     this.period = settings.period();
-    this.maximumMainResidentSize = (int) (settings.maximumSize() * settings.percentMain());
-    this.maximumFilterSize = settings.maximumSize() - maximumMainResidentSize;
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
+    this.maximumMainResidentSize = (int) (maximumSize * settings.percentMain());
+    this.maximumFilterSize = maximumSize - maximumMainResidentSize;
     this.policyStats = new PolicyStats("irr.AdaptiveFrd");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.headFilter = new Node();
     this.headMain = new Node();
     this.indicator = new Indicator(config);

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/irr/LirsPolicy.java

@@ -27,6 +27,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -78,11 +79,11 @@ public final class LirsPolicy implements KeyOnlyPolicy {
 
   public LirsPolicy(Config config) {
     LirsSettings settings = new LirsSettings(config);
-    this.maximumNonResidentSize = (int) (settings.maximumSize() * settings.nonResidentMultiplier());
-    this.maximumHotSize = (int) (settings.maximumSize() * settings.percentHot());
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
+    this.maximumNonResidentSize = (int) (maximumSize * settings.nonResidentMultiplier());
+    this.maximumHotSize = (int) (maximumSize * settings.percentHot());
     this.policyStats = new PolicyStats("irr.Lirs");
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.evicted = new ArrayList<>();
     this.headNR = new Node();
     this.headS = new Node();

simulator/src/main/java/com/github/benmanes/caffeine/cache/simulator/policy/linked/FrequentlyUsedPolicy.java

@@ -30,6 +30,7 @@
 import com.github.benmanes.caffeine.cache.simulator.policy.Policy.KeyOnlyPolicy;
 import com.github.benmanes.caffeine.cache.simulator.policy.PolicyStats;
 import com.google.common.base.MoreObjects;
+import com.google.common.primitives.Ints;
 import com.typesafe.config.Config;
 
 import it.unimi.dsi.fastutil.longs.Long2ObjectMap;
@@ -50,11 +51,11 @@ public final class FrequentlyUsedPolicy implements KeyOnlyPolicy {
   final int maximumSize;
 
   public FrequentlyUsedPolicy(Admission admission, EvictionPolicy policy, Config config) {
+    BasicSettings settings = new BasicSettings(config);
     this.policyStats = new PolicyStats(admission.format("linked." + policy.label()));
+    this.maximumSize = Ints.checkedCast(settings.maximumSize());
     this.admittor = admission.from(config, policyStats);
-    BasicSettings settings = new BasicSettings(config);
     this.data = new Long2ObjectOpenHashMap<>();
-    this.maximumSize = settings.maximumSize();
     this.policy = requireNonNull(policy);
     this.freq0 = new FrequencyNode();
   }