- Use xxh3 hash for better hashing, and cache the hash value.

2 files changed, 32 insertions, 6 deletions

diff --git a/vespajlib/src/main/java/com/yahoo/compress/Hasher.java b/vespajlib/src/main/java/com/yahoo/compress/Hasher.java
index 92a9ed26085..d991a98a5d0 100644
--- a/vespajlib/src/main/java/com/yahoo/compress/Hasher.java
+++ b/vespajlib/src/main/java/com/yahoo/compress/Hasher.java
@@ -8,8 +8,25 @@ import net.openhft.hashing.LongHashFunction;
  * @author baldersheim
  */
 public class Hasher {
+    private final LongHashFunction hasher;
     /** Uses net.openhft.hashing.LongHashFunction.xx3() */
     public static long xxh3(byte [] data) {
         return LongHashFunction.xx3().hashBytes(data);
     }
+    public static long xxh3(byte [] data, long seed) {
+        return LongHashFunction.xx3(seed).hashBytes(data);
+    }
+
+    private Hasher(LongHashFunction hasher) {
+        this.hasher = hasher;
+    }
+    public static Hasher of(long seed) {
+        return new Hasher(LongHashFunction.xx3(seed));
+    }
+    public long hash(long v) {
+        return hasher.hashLong(v);
+    }
+    public long hash(String s) {
+        return hasher.hashChars(s);
+    }
 }
diff --git a/vespajlib/src/main/java/com/yahoo/tensor/TensorAddress.java b/vespajlib/src/main/java/com/yahoo/tensor/TensorAddress.java
index f841b7757fb..e3dfc297da6 100644
--- a/vespajlib/src/main/java/com/yahoo/tensor/TensorAddress.java
+++ b/vespajlib/src/main/java/com/yahoo/tensor/TensorAddress.java
@@ -1,6 +1,7 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 package com.yahoo.tensor;
 
+import com.yahoo.compress.Hasher;
 import com.yahoo.tensor.impl.NumericTensorAddress;
 import com.yahoo.tensor.impl.StringTensorAddress;
 
@@ -28,6 +29,9 @@ public abstract class TensorAddress implements Comparable<TensorAddress> {
         return NumericTensorAddress.of(labels);
     }
 
+    // If the hash ends up as 0 it will be recalculated everytime, but that is not a practical issue.
+    private int cached_hash = 0;
+
     /** Returns the number of labels in this */
     public abstract int size();
 
@@ -62,12 +66,15 @@ public abstract class TensorAddress implements Comparable<TensorAddress> {
 
     @Override
     public int hashCode() {
-        int result = 1;
+        if (cached_hash != 0) return cached_hash;
+        Hasher hasher = Hasher.of(0);
+        long hash = 0;
         for (int i = 0; i < size(); i++) {
-            if (label(i) != null)
-                result = 31 * result + label(i).hashCode();
+            hash = hash ^ hasher.hash(label(i));
         }
-        return result;
+        int low = (int) hash;
+        int high = (int) (hash >> 32);
+        return cached_hash = low ^ high;
     }
 
     @Override
@@ -105,7 +112,7 @@ public abstract class TensorAddress implements Comparable<TensorAddress> {
     public static class Builder {
 
         final TensorType type;
-        final String[] labels;
+        String[] labels;
 
         public Builder(TensorType type) {
             this(type, new String[type.dimensions().size()]);
@@ -162,7 +169,9 @@ public abstract class TensorAddress implements Comparable<TensorAddress> {
 
         public TensorAddress build() {
             validate();
-            return TensorAddress.of(labels);
+            var address = StringTensorAddress.unsafeOf(labels);
+            labels = null;
+            return address;
         }
 
     }