wip

simdjson · Nov 10, 2023 · bff19f3 · bff19f3
1 parent 32d9059
commit bff19f3
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 8 deletions.
diff --git a/src/main/java/org/simdjson/StructuralIndexer.java b/src/main/java/org/simdjson/StructuralIndexer.java
@@ -1,22 +1,33 @@
 package org.simdjson;
 
 import jdk.incubator.vector.ByteVector;
+import jdk.incubator.vector.IntVector;
 import jdk.incubator.vector.VectorSpecies;
 import java.lang.invoke.MethodType;
 
 import static jdk.incubator.vector.VectorOperators.UNSIGNED_LE;
 
 class StructuralIndexer {
 
+    static final VectorSpecies<Integer> INT_SPECIES;
     static final VectorSpecies<Byte> SPECIES;
     static final int N_CHUNKS;
 
     static {
         String species = System.getProperty("org.simdjson.species", "preferred");
-        SPECIES = switch(species) {
-            case "preferred" -> ByteVector.SPECIES_PREFERRED;
-            case "512" -> ByteVector.SPECIES_512;      
-            case "256" -> ByteVector.SPECIES_256;
+        switch(species) {
+            case "preferred" -> {
+                SPECIES = ByteVector.SPECIES_PREFERRED;
+                INT_SPECIES = IntVector.SPECIES_PREFERRED;
+            }
+            case "512" -> {
+                SPECIES = ByteVector.SPECIES_512;
+                INT_SPECIES = IntVector.SPECIES_512;
+            }
+            case "256" -> {
+                SPECIES = ByteVector.SPECIES_256;
+                INT_SPECIES = IntVector.SPECIES_256;
+            }
             default -> throw new IllegalArgumentException("Unsupported vector species: " + species);
         };              
         N_CHUNKS = 64 / SPECIES.vectorByteSize();

diff --git a/src/main/java/org/simdjson/Utf8Validator.java b/src/main/java/org/simdjson/Utf8Validator.java
@@ -5,10 +5,10 @@
 import java.util.Arrays;
 
 public class Utf8Validator {
-    private static final VectorSpecies<Byte> VECTOR_SPECIES = ByteVector.SPECIES_256;
+    private static final VectorSpecies<Byte> VECTOR_SPECIES = StructuralIndexer.SPECIES;
     private static final ByteVector INCOMPLETE_CHECK = getIncompleteCheck();
-    private static final VectorShuffle<Integer> SHIFT_FOUR_BYTES_FORWARD = VectorShuffle.iota(IntVector.SPECIES_256,
-            IntVector.SPECIES_256.elementSize() - 1, 1, true);
+    private static final VectorShuffle<Integer> SHIFT_FOUR_BYTES_FORWARD = VectorShuffle.iota(StructuralIndexer.INT_SPECIES,
+            StructuralIndexer.INT_SPECIES.elementSize() - 1, 1, true);
     private static final ByteVector LOW_NIBBLE_MASK = ByteVector.broadcast(VECTOR_SPECIES, 0b0000_1111);
     private static final ByteVector ALL_ASCII_MASK = ByteVector.broadcast(VECTOR_SPECIES, (byte) 0b1000_0000);
 
@@ -40,7 +40,7 @@ public static void validate(byte[] inputBytes) {
 
                 errors |= secondCheck.compare(VectorOperators.NE, 0).toLong();
             }
-            previousFourUtf8Bytes = utf8Vector.reinterpretAsInts().lane(IntVector.SPECIES_256.length() - 1);
+            previousFourUtf8Bytes = utf8Vector.reinterpretAsInts().lane(StructuralIndexer.INT_SPECIES.length() - 1);
         }
 
         // if the input file doesn't align with the vector width, pad the missing bytes with zero