[WIP] Initial SIMD support

aot/src/main/java/com/dylibso/chicory/aot/AotMachine.java

@@ -369,6 +369,12 @@ public Value[] call(int funcId, Value[] args) throws ChicoryException {
         }
     }
 
+    // TODO: implement SIMD support
+    @Override
+    public Map<OpCode, OpImpl> additionalOpCodes() {
+        return Map.of();
+    }
+
     private MethodHandle[] compile(Class<?> clazz) {
         var functions = module.functionSection();
         var compiled = new MethodHandle[functionImports + functions.functionCount()];

pom.xml

@@ -45,6 +45,7 @@
     <module>log</module>
     <module>runtime</module>
     <module>runtime-tests</module>
+    <module>simd</module>
     <module>test-gen-plugin</module>
     <module>wabt</module>
     <module>wasi</module>
@@ -162,6 +163,11 @@
         <artifactId>runtime</artifactId>
         <version>${project.version}</version>
       </dependency>
+      <dependency>
+        <groupId>com.dylibso.chicory</groupId>
+        <artifactId>simd</artifactId>
+        <version>${project.version}</version>
+      </dependency>
       <dependency>
         <groupId>com.dylibso.chicory</groupId>
         <artifactId>wabt</artifactId>

runtime-tests/pom.xml

@@ -13,12 +13,21 @@
   <name>Chicory - Runtime tests</name>
   <description>Tests for the Chicory runtime using the WebAssembly testsuite</description>
 
+  <properties>
+    <maven.compiler.release>21</maven.compiler.release>
+  </properties>
+
   <dependencies>
     <dependency>
       <groupId>com.dylibso.chicory</groupId>
       <artifactId>runtime</artifactId>
       <scope>test</scope>
     </dependency>
+    <dependency>
+      <groupId>com.dylibso.chicory</groupId>
+      <artifactId>simd</artifactId>
+      <scope>test</scope>
+    </dependency>
     <dependency>
       <groupId>com.dylibso.chicory</groupId>
       <artifactId>wabt</artifactId>
@@ -48,6 +57,7 @@
 
   <build>
     <plugins>
+
       <plugin>
         <groupId>com.dylibso.chicory</groupId>
         <artifactId>test-gen-plugin</artifactId>
@@ -118,6 +128,7 @@
             <wast>ref_null.wast</wast>
             <wast>return.wast</wast>
             <wast>select.wast</wast>
+            <wast>simd_load.wast</wast>
             <wast>skip-stack-guard-page.wast</wast>
             <wast>stack.wast</wast>
             <wast>start.wast</wast>
@@ -194,7 +205,6 @@
             <wast>simd_int_to_int_extend.wast</wast>
             <wast>simd_lane.wast</wast>
             <wast>simd_linking.wast</wast>
-            <wast>simd_load.wast</wast>
             <wast>simd_load16_lane.wast</wast>
             <wast>simd_load32_lane.wast</wast>
             <wast>simd_load64_lane.wast</wast>
@@ -218,6 +228,20 @@
           </execution>
         </executions>
       </plugin>
+      <plugin>
+        <groupId>org.apache.maven.plugins</groupId>
+        <artifactId>maven-failsafe-plugin</artifactId>
+        <configuration>
+          <argLine>--add-modules jdk.incubator.vector</argLine>
+        </configuration>
+      </plugin>
+      <plugin>
+        <groupId>org.apache.maven.plugins</groupId>
+        <artifactId>maven-surefire-plugin</artifactId>
+        <configuration>
+          <argLine>--add-modules jdk.incubator.vector</argLine>
+        </configuration>
+      </plugin>
     </plugins>
   </build>
 </project>

runtime-tests/src/test/java/com/dylibso/chicory/testing/TestModule.java

@@ -2,7 +2,9 @@
 
 import com.dylibso.chicory.runtime.HostImports;
 import com.dylibso.chicory.runtime.Instance;
+import com.dylibso.chicory.runtime.InterpreterMachine;
 import com.dylibso.chicory.runtime.Store;
+import com.dylibso.chicory.simd.Simd;
 import com.dylibso.chicory.wabt.Wat2Wasm;
 import com.dylibso.chicory.wasm.Module;
 import com.dylibso.chicory.wasm.Parser;
@@ -62,6 +64,9 @@ public TestModule(Module module) {
 
     public Instance instantiate(Store s) {
         HostImports hostImports = s.toHostImports();
-        return Instance.builder(module).withHostImports(hostImports).build();
+        return Instance.builder(module)
+                .withMachineFactory((inst) -> new InterpreterMachine(inst, Simd.opcodesImpl))
+                .withHostImports(hostImports)
+                .build();
     }
 }

runtime/src/main/java/com/dylibso/chicory/runtime/InterpreterMachine.java

@@ -12,22 +12,35 @@
 import java.util.ArrayDeque;
 import java.util.Deque;
 import java.util.List;
+import java.util.Map;
 
 /**
  * This is responsible for holding and interpreting the Wasm code.
  */
-class InterpreterMachine implements Machine {
+public class InterpreterMachine implements Machine {
 
     private final MStack stack;
 
     private final Deque<StackFrame> callStack;
 
     private final Instance instance;
 
+    private final Map<OpCode, OpImpl> additionalOpcodes;
+
     public InterpreterMachine(Instance instance) {
+        this(instance, Map.of());
+    }
+
+    public InterpreterMachine(Instance instance, Map<OpCode, OpImpl> additionalOpcodes) {
         this.instance = instance;
         stack = new MStack();
         this.callStack = new ArrayDeque<>();
+        this.additionalOpcodes = additionalOpcodes;
+    }
+
+    @Override
+    public Map<OpCode, OpImpl> additionalOpCodes() {
+        return additionalOpcodes;
     }
 
     @Override
@@ -57,7 +70,7 @@ public static Value[] call(
         if (func != null) {
             var stackFrame =
                     new StackFrame(func.instructions(), instance, funcId, args, func.localTypes());
-            stackFrame.pushCtrl(OpCode.CALL, 0, type.returns().size(), stack.size());
+            stackFrame.pushCtrl(OpCode.CALL, 0, sizeOf(type.returns()), stack.size());
             callStack.push(stackFrame);
 
             try {
@@ -67,7 +80,7 @@ public static Value[] call(
             }
         } else {
             var stackFrame = new StackFrame(instance, funcId, args, List.of());
-            stackFrame.pushCtrl(OpCode.CALL, 0, type.returns().size(), stack.size());
+            stackFrame.pushCtrl(OpCode.CALL, 0, sizeOf(type.returns()), stack.size());
             callStack.push(stackFrame);
 
             var results = instance.callHostFunction(funcId, args);
@@ -95,7 +108,7 @@ public static Value[] call(
             return null;
         }
 
-        var totalResults = type.returns().size();
+        var totalResults = sizeOf(type.returns());
         var results = new Value[totalResults];
         for (var i = totalResults - 1; i >= 0; i--) {
             results[i] = stack.pop();
@@ -745,8 +758,15 @@ static void eval(MStack stack, Instance instance, Deque<StackFrame> callStack)
                     ELEM_DROP(instance, operands);
                     break;
                 default:
-                    throw new RuntimeException(
-                            "Machine doesn't recognize Instruction " + instruction);
+                    {
+                        var additionalOpcodes = instance.getMachine().additionalOpCodes();
+                        if (additionalOpcodes.containsKey(opcode)) {
+                            additionalOpcodes.get(opcode).invoke(stack, instance, operands);
+                        } else {
+                            throw new RuntimeException(
+                                    "Machine doesn't recognize Instruction " + instruction);
+                        }
+                    }
             }
         }
     }
@@ -1653,7 +1673,7 @@ private static void MEMORY_GROW(MStack stack, Instance instance) {
         stack.push(Value.i32(nPages));
     }
 
-    private static int readMemPtr(MStack stack, Operands operands) {
+    public static int readMemPtr(MStack stack, Operands operands) {
         int offset = stack.pop().asInt();
         if (operands.get(1) < 0 || operands.get(1) >= Integer.MAX_VALUE || offset < 0) {
             throw new WASMRuntimeException("out of bounds memory access");
@@ -1861,7 +1881,7 @@ private static int numberOfParams(Instance instance, AnnotatedInstruction scope)
         if (ValueType.isValid(typeId)) {
             return 0;
         }
-        return instance.type(typeId).params().size();
+        return sizeOf(instance.type(typeId).params());
     }
 
     private static int numberOfValuesToReturn(Instance instance, AnnotatedInstruction scope) {
@@ -1873,9 +1893,25 @@ private static int numberOfValuesToReturn(Instance instance, AnnotatedInstructio
             return 0;
         }
         if (ValueType.isValid(typeId)) {
-            return 1;
+            if (ValueType.forId(typeId).isVec()) {
+                return 2;
+            } else {
+                return 1;
+            }
+        }
+        return sizeOf(instance.type(typeId).returns());
+    }
+
+    private static int sizeOf(List<ValueType> args) {
+        int total = 0;
+        for (var a : args) {
+            if (a.isVec()) {
+                total += 2;
+            } else {
+                total += 1;
+            }
         }
-        return instance.type(typeId).returns().size();
+        return total;
     }
 
     private static void BLOCK(
@@ -1941,7 +1977,7 @@ static Value[] extractArgsForParams(MStack stack, List<ValueType> params) {
         if (params == null) {
             return Value.EMPTY_VALUES;
         }
-        var args = new Value[params.size()];
+        var args = new Value[sizeOf(params)];
         for (var i = params.size(); i > 0; i--) {
             var p = stack.pop();
             var t = params.get(i - 1);
@@ -1984,7 +2020,7 @@ private static void checkInterruption() {
     }
 
     @FunctionalInterface
-    private interface Operands {
+    public interface Operands {
         long get(int index);
     }
 }

runtime/src/main/java/com/dylibso/chicory/runtime/Machine.java

@@ -1,9 +1,18 @@
 package com.dylibso.chicory.runtime;
 
 import com.dylibso.chicory.wasm.exceptions.ChicoryException;
+import com.dylibso.chicory.wasm.types.OpCode;
 import com.dylibso.chicory.wasm.types.Value;
+import java.util.Map;
 
 public interface Machine {
 
     Value[] call(int funcId, Value[] args) throws ChicoryException;
+
+    Map<OpCode, OpImpl> additionalOpCodes();
+
+    @FunctionalInterface
+    public interface OpImpl {
+        void invoke(MStack stack, Instance instance, InterpreterMachine.Operands operands);
+    }
 }

runtime/src/main/java/com/dylibso/chicory/runtime/Memory.java

@@ -364,6 +364,14 @@ public Value readF64(int addr) {
         }
     }
 
+    public Value readV128(int addr) {
+        try {
+            return Value.v128(buffer.getLong(addr));
+        } catch (IndexOutOfBoundsException e) {
+            throw new WASMRuntimeException("out of bounds memory access");
+        }
+    }
+
     public void zero() {
         this.fill((byte) 0);
     }

simd/pom.xml

@@ -0,0 +1,69 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+  <modelVersion>4.0.0</modelVersion>
+
+  <parent>
+    <groupId>com.dylibso.chicory</groupId>
+    <artifactId>chicory</artifactId>
+    <version>999-SNAPSHOT</version>
+    <relativePath>../pom.xml</relativePath>
+  </parent>
+  <artifactId>simd</artifactId>
+  <packaging>jar</packaging>
+  <name>Chicory - SIMD</name>
+  <description>SIMD instructions support for Chicory</description>
+
+  <properties>
+    <maven.compiler.release>21</maven.compiler.release>
+
+    <!-- TODO: fixme working around "[WARNING] using incubating module(s): jdk.incubator.vector" -->
+    <maven.compiler.failOnWarning>false</maven.compiler.failOnWarning>
+  </properties>
+
+  <dependencies>
+    <dependency>
+      <groupId>com.dylibso.chicory</groupId>
+      <artifactId>runtime</artifactId>
+    </dependency>
+    <dependency>
+      <groupId>com.dylibso.chicory</groupId>
+      <artifactId>wasm</artifactId>
+    </dependency>
+    <dependency>
+      <groupId>org.junit.jupiter</groupId>
+      <artifactId>junit-jupiter-api</artifactId>
+      <scope>test</scope>
+    </dependency>
+    <dependency>
+      <groupId>org.junit.jupiter</groupId>
+      <artifactId>junit-jupiter-engine</artifactId>
+      <scope>test</scope>
+    </dependency>
+  </dependencies>
+
+  <build>
+    <plugins>
+      <plugin>
+        <groupId>org.apache.maven.plugins</groupId>
+        <artifactId>maven-compiler-plugin</artifactId>
+        <configuration>
+          <compilerArgs>
+            <arg>--add-modules</arg>
+            <arg>jdk.incubator.vector</arg>
+          </compilerArgs>
+        </configuration>
+      </plugin>
+      <plugin>
+        <groupId>org.apache.maven.plugins</groupId>
+        <artifactId>maven-javadoc-plugin</artifactId>
+        <configuration>
+          <!-- Workaround: "MavenReportException: Error while generating Javadoc:" -->
+          <failOnError>false</failOnError>
+          <failOnWarnings>false</failOnWarnings>
+          <additionalJOption>-Xdoclint:none</additionalJOption>
+        </configuration>
+      </plugin>
+    </plugins>
+  </build>
+
+</project>