diff --git a/src/hotspot/share/classfile/vmIntrinsics.hpp b/src/hotspot/share/classfile/vmIntrinsics.hpp
index 18300ca3465..78193f2d0b3 100644
--- a/src/hotspot/share/classfile/vmIntrinsics.hpp
+++ b/src/hotspot/share/classfile/vmIntrinsics.hpp
@@ -858,14 +858,15 @@ class methodHandle;
                                       "Ljava/lang/Object;ILjdk/internal/vm/vector/VectorSupport$StoreVectorOperation;)V")                      \
    do_name(vector_store_op_name,     "store")                                                                                                  \
                                                                                                                                                \
-  do_intrinsic(_VectorStoreMaskedOp, jdk_internal_vm_vector_VectorSupport, vector_store_masked_op_name, vector_store_masked_op_sig, F_S)             \
-   do_signature(vector_store_masked_op_sig, "(Ljava/lang/Class;Ljava/lang/Class;Ljava/lang/Class;ILjava/lang/Object;JLjdk/internal/vm/vector/VectorSupport$Vector;"    \
+  do_intrinsic(_VectorStoreMaskedOp, jdk_internal_vm_vector_VectorSupport, vector_store_masked_op_name, vector_store_masked_op_sig, F_S)       \
+   do_signature(vector_store_masked_op_sig, "(Ljava/lang/Class;Ljava/lang/Class;Ljava/lang/Class;ILjava/lang/Object;JLjdk/internal/vm/vector/VectorSupport$Vector;"  \
                                              "Ljdk/internal/vm/vector/VectorSupport$VectorMask;Ljava/lang/Object;I"                            \
                                              "Ljdk/internal/vm/vector/VectorSupport$StoreVectorMaskedOperation;)V")                            \
    do_name(vector_store_masked_op_name,     "storeMasked")                                                                                     \
                                                                                                                                                \
-  do_intrinsic(_VectorReductionCoerced, jdk_internal_vm_vector_VectorSupport, vector_reduction_coerced_name, vector_reduction_coerced_sig, F_S) \
-   do_signature(vector_reduction_coerced_sig, "(ILjava/lang/Class;Ljava/lang/Class;ILjdk/internal/vm/vector/VectorSupport$Vector;Ljava/util/function/Function;)J") \
+  do_intrinsic(_VectorReductionCoerced, jdk_internal_vm_vector_VectorSupport, vector_reduction_coerced_name, vector_reduction_coerced_sig, F_S)\
+   do_signature(vector_reduction_coerced_sig, "(ILjava/lang/Class;Ljava/lang/Class;Ljava/lang/Class;ILjava/lang/Object;Ljava/lang/Object;"     \
+                                               "Ljdk/internal/vm/vector/VectorSupport$ReductionOperation;)J")                                  \
    do_name(vector_reduction_coerced_name, "reductionCoerced")                                                                                  \
                                                                                                                                                \
   do_intrinsic(_VectorTest, jdk_internal_vm_vector_VectorSupport, vector_test_name, vector_test_sig, F_S)                                      \
diff --git a/src/hotspot/share/opto/vectorIntrinsics.cpp b/src/hotspot/share/opto/vectorIntrinsics.cpp
index 5a914a4d60c..d83afca8595 100644
--- a/src/hotspot/share/opto/vectorIntrinsics.cpp
+++ b/src/hotspot/share/opto/vectorIntrinsics.cpp
@@ -1270,16 +1270,17 @@ bool LibraryCallKit::inline_vector_gather_scatter(bool is_scatter) {
   return true;
 }
 
-// <V extends Vector<?,?>>
-// long reductionCoerced(int oprId, Class<?> vectorClass, Class<?> elementType, int vlen,
-//                       V v,
-//                       Function<V,Long> defaultImpl)
-
+// <V, M>
+// long reductionCoerced(int oprId, Class<? extends V> vectorClass, Class<? extends M> maskClass,
+//                       Class<?> elementType, int length, V v, M m,
+//                       ReductionOperation<V, M> defaultImpl) {
+//
 bool LibraryCallKit::inline_vector_reduction() {
   const TypeInt*     opr          = gvn().type(argument(0))->isa_int();
   const TypeInstPtr* vector_klass = gvn().type(argument(1))->isa_instptr();
-  const TypeInstPtr* elem_klass   = gvn().type(argument(2))->isa_instptr();
-  const TypeInt*     vlen         = gvn().type(argument(3))->isa_int();
+  const TypeInstPtr* mask_klass   = gvn().type(argument(2))->isa_instptr();
+  const TypeInstPtr* elem_klass   = gvn().type(argument(3))->isa_instptr();
+  const TypeInt*     vlen         = gvn().type(argument(4))->isa_int();
 
   if (opr == NULL || vector_klass == NULL || elem_klass == NULL || vlen == NULL ||
       !opr->is_con() || vector_klass->const_oop() == NULL || elem_klass->const_oop() == NULL || !vlen->is_con()) {
@@ -1287,8 +1288,8 @@ bool LibraryCallKit::inline_vector_reduction() {
       tty->print_cr("  ** missing constant: opr=%s vclass=%s etype=%s vlen=%s",
                     NodeClassNames[argument(0)->Opcode()],
                     NodeClassNames[argument(1)->Opcode()],
-                    NodeClassNames[argument(2)->Opcode()],
-                    NodeClassNames[argument(3)->Opcode()]);
+                    NodeClassNames[argument(3)->Opcode()],
+                    NodeClassNames[argument(4)->Opcode()]);
     }
     return false; // not enough info for intrinsification
   }
@@ -1305,16 +1306,51 @@ bool LibraryCallKit::inline_vector_reduction() {
     }
     return false; // should be primitive type
   }
+
+  const Type* vmask_type = gvn().type(argument(6));
+  bool is_masked_op = vmask_type != TypePtr::NULL_PTR;
+  if (is_masked_op) {
+    if (mask_klass == NULL || mask_klass->const_oop() == NULL) {
+      if (C->print_intrinsics()) {
+        tty->print_cr("  ** missing constant: maskclass=%s", NodeClassNames[argument(2)->Opcode()]);
+      }
+      return false; // not enough info for intrinsification
+    }
+
+    if (!is_klass_initialized(mask_klass)) {
+      if (C->print_intrinsics()) {
+        tty->print_cr("  ** mask klass argument not initialized");
+      }
+      return false;
+    }
+
+    if (vmask_type->maybe_null()) {
+      if (C->print_intrinsics()) {
+        tty->print_cr("  ** null mask values are not allowed for masked op");
+      }
+      return false;
+    }
+  }
+
   BasicType elem_bt = elem_type->basic_type();
   int num_elem = vlen->get_con();
-
   int opc  = VectorSupport::vop2ideal(opr->get_con(), elem_bt);
   int sopc = ReductionNode::opcode(opc, elem_bt);
 
-  // TODO When mask usage is supported, VecMaskNotUsed needs to be VecMaskUseLoad.
-  if (!arch_supports_vector(sopc, num_elem, elem_bt, VecMaskNotUsed)) {
+  // When using mask, mask use type needs to be VecMaskUseLoad.
+  if (!arch_supports_vector(sopc, num_elem, elem_bt, is_masked_op ? VecMaskUseLoad : VecMaskNotUsed)) {
+    if (C->print_intrinsics()) {
+      tty->print_cr("  ** not supported: arity=1 op=%d/reduce vlen=%d etype=%s is_masked_op=%d",
+                    sopc, num_elem, type2name(elem_bt), is_masked_op ? 1 : 0);
+    }
+    return false;
+  }
+
+  // Return true if current platform has implemented the masked operation with predicate feature.
+  bool use_predicate = is_masked_op && arch_supports_vector(sopc, num_elem, elem_bt, VecMaskUsePred);
+  if (is_masked_op && !use_predicate && !arch_supports_vector(Op_VectorBlend, num_elem, elem_bt, VecMaskUseLoad)) {
     if (C->print_intrinsics()) {
-      tty->print_cr("  ** not supported: arity=1 op=%d/reduce vlen=%d etype=%s ismask=no",
+      tty->print_cr("  ** not supported: arity=1 op=%d/reduce vlen=%d etype=%s is_masked_op=1",
                     sopc, num_elem, type2name(elem_bt));
     }
     return false;
@@ -1323,33 +1359,64 @@ bool LibraryCallKit::inline_vector_reduction() {
   ciKlass* vbox_klass = vector_klass->const_oop()->as_instance()->java_lang_Class_klass();
   const TypeInstPtr* vbox_type = TypeInstPtr::make_exact(TypePtr::NotNull, vbox_klass);
 
-  Node* opd = unbox_vector(argument(4), vbox_type, elem_bt, num_elem);
+  Node* opd = unbox_vector(argument(5), vbox_type, elem_bt, num_elem);
   if (opd == NULL) {
     return false; // operand unboxing failed
   }
 
+  Node* mask = NULL;
+  if (is_masked_op) {
+    ciKlass* mbox_klass = mask_klass->const_oop()->as_instance()->java_lang_Class_klass();
+    assert(is_vector_mask(mbox_klass), "argument(2) should be a mask class");
+    const TypeInstPtr* mbox_type = TypeInstPtr::make_exact(TypePtr::NotNull, mbox_klass);
+    mask = unbox_vector(argument(6), mbox_type, elem_bt, num_elem);
+    if (mask == NULL) {
+      if (C->print_intrinsics()) {
+        tty->print_cr("  ** unbox failed mask=%s",
+                      NodeClassNames[argument(6)->Opcode()]);
+      }
+      return false;
+    }
+  }
+
   Node* init = ReductionNode::make_reduction_input(gvn(), opc, elem_bt);
-  Node* rn = gvn().transform(ReductionNode::make(opc, NULL, init, opd, elem_bt));
+  Node* value = NULL;
+  if (mask == NULL) {
+    assert(!is_masked_op, "Masked op needs the mask value never null");
+    value = ReductionNode::make(opc, NULL, init, opd, elem_bt);
+  } else {
+    if (use_predicate) {
+      if (C->print_intrinsics()) {
+        tty->print_cr("  ** predicate feature is not supported on current platform!");
+      }
+      return false;
+    } else {
+      Node* reduce_identity = gvn().transform(VectorNode::scalar2vector(init, num_elem, Type::get_const_basic_type(elem_bt)));
+      value = gvn().transform(new VectorBlendNode(reduce_identity, opd, mask));
+      value = ReductionNode::make(opc, NULL, init, value, elem_bt);
+    }
+  }
+  value = gvn().transform(value);
 
   Node* bits = NULL;
   switch (elem_bt) {
     case T_BYTE:
     case T_SHORT:
     case T_INT: {
-      bits = gvn().transform(new ConvI2LNode(rn));
+      bits = gvn().transform(new ConvI2LNode(value));
       break;
     }
     case T_FLOAT: {
-      rn   = gvn().transform(new MoveF2INode(rn));
-      bits = gvn().transform(new ConvI2LNode(rn));
+      value = gvn().transform(new MoveF2INode(value));
+      bits  = gvn().transform(new ConvI2LNode(value));
       break;
     }
     case T_DOUBLE: {
-      bits = gvn().transform(new MoveD2LNode(rn));
+      bits = gvn().transform(new MoveD2LNode(value));
       break;
     }
     case T_LONG: {
-      bits = rn; // no conversion needed
+      bits = value; // no conversion needed
       break;
     }
     default: fatal("%s", type2name(elem_bt));
diff --git a/src/hotspot/share/opto/vectornode.cpp b/src/hotspot/share/opto/vectornode.cpp
index b51b0a51680..e955bfc5367 100644
--- a/src/hotspot/share/opto/vectornode.cpp
+++ b/src/hotspot/share/opto/vectornode.cpp
@@ -1076,7 +1076,9 @@ Node* ReductionNode::make_reduction_input(PhaseGVN& gvn, int opc, BasicType bt)
     case Op_MinReductionV:
       switch (bt) {
         case T_BYTE:
+          return gvn.makecon(TypeInt::make(max_jbyte));
         case T_SHORT:
+          return gvn.makecon(TypeInt::make(max_jshort));
         case T_INT:
           return gvn.makecon(TypeInt::MAX);
         case T_LONG:
@@ -1091,7 +1093,9 @@ Node* ReductionNode::make_reduction_input(PhaseGVN& gvn, int opc, BasicType bt)
     case Op_MaxReductionV:
       switch (bt) {
         case T_BYTE:
+          return gvn.makecon(TypeInt::make(min_jbyte));
         case T_SHORT:
+          return gvn.makecon(TypeInt::make(min_jshort));
         case T_INT:
           return gvn.makecon(TypeInt::MIN);
         case T_LONG:
diff --git a/src/java.base/share/classes/jdk/internal/vm/vector/VectorSupport.java b/src/java.base/share/classes/jdk/internal/vm/vector/VectorSupport.java
index 626a264f5a9..15a1cdac56b 100644
--- a/src/java.base/share/classes/jdk/internal/vm/vector/VectorSupport.java
+++ b/src/java.base/share/classes/jdk/internal/vm/vector/VectorSupport.java
@@ -212,14 +212,19 @@ V indexVector(Class<? extends V> vClass, Class<E> E, int length,
 
     @IntrinsicCandidate
     public static
-    <V extends Vector<?>>
-    long reductionCoerced(int oprId, Class<?> vectorClass, Class<?> elementType, int length,
-                          V v,
-                          Function<V,Long> defaultImpl) {
+    <V, M>
+    long reductionCoerced(int oprId, Class<? extends V> vectorClass, Class<? extends M> maskClass,
+                          Class<?> elementType, int length, V v, M m,
+                          ReductionOperation<V, M> defaultImpl) {
         assert isNonCapturingLambda(defaultImpl) : defaultImpl;
-        return defaultImpl.apply(v);
+        return defaultImpl.apply(v, m);
     }
 
+    public interface ReductionOperation<V, M> {
+        long apply(V v, M mask);
+    }
+
+
     /* ============================================================================ */
 
     public interface VecExtractOp<V> {
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte128Vector.java
index 1407b6d8c67..845379eaffa 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte128Vector.java
@@ -236,8 +236,8 @@ Byte128Vector tOp(Vector<Byte> v1, Vector<Byte> v2,
 
     @ForceInline
     final @Override
-    byte rOp(byte v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final byte reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final byte reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Byte> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Byte128Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Byte> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Byte128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte256Vector.java
index 75c67bb9035..a06f5ca518b 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte256Vector.java
@@ -236,8 +236,8 @@ Byte256Vector tOp(Vector<Byte> v1, Vector<Byte> v2,
 
     @ForceInline
     final @Override
-    byte rOp(byte v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final byte reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final byte reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Byte> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Byte256Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Byte> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Byte256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte512Vector.java
index 433558cf4fd..ba5b303768c 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte512Vector.java
@@ -236,8 +236,8 @@ Byte512Vector tOp(Vector<Byte> v1, Vector<Byte> v2,
 
     @ForceInline
     final @Override
-    byte rOp(byte v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final byte reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final byte reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Byte> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Byte512Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Byte> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Byte512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte64Vector.java
index 9238934838b..3b30a8d5f3f 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Byte64Vector.java
@@ -236,8 +236,8 @@ Byte64Vector tOp(Vector<Byte> v1, Vector<Byte> v2,
 
     @ForceInline
     final @Override
-    byte rOp(byte v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final byte reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final byte reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Byte> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Byte64Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Byte> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Byte64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteMaxVector.java
index e7344d198f5..1382e4a8888 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteMaxVector.java
@@ -236,8 +236,8 @@ ByteMaxVector tOp(Vector<Byte> v1, Vector<Byte> v2,
 
     @ForceInline
     final @Override
-    byte rOp(byte v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final byte reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final byte reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Byte> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, ByteMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Byte> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, ByteMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteVector.java
index 3d8cc1b46b3..66676e1faa8 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ByteVector.java
@@ -287,7 +287,22 @@ ByteVector tOpTemplate(Vector<Byte> o1,
 
     /*package-private*/
     abstract
-    byte rOp(byte v, FBinOp f);
+    byte rOp(byte v, VectorMask<Byte> m, FBinOp f);
+
+    @ForceInline
+    final
+    byte rOpTemplate(byte v, VectorMask<Byte> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        byte[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Byte>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     byte rOpTemplate(byte v, FBinOp f) {
@@ -2532,9 +2547,18 @@ public abstract byte reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     byte reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Byte>> maskClass,
                                VectorMask<Byte> m) {
-        ByteVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            ByteVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, byte.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, ByteVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2549,30 +2573,34 @@ byte reduceLanesTemplate(VectorOperators.Associative op) {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), byte.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((byte)0, (i, a, b) -> (byte)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((byte)1, (i, a, b) -> (byte)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (byte) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (byte) Math.max(a, b)));
-              case VECTOR_OP_AND: return v ->
-                      toBits(v.rOp((byte)-1, (i, a, b) -> (byte)(a & b)));
-              case VECTOR_OP_OR: return v ->
-                      toBits(v.rOp((byte)0, (i, a, b) -> (byte)(a | b)));
-              case VECTOR_OP_XOR: return v ->
-                      toBits(v.rOp((byte)0, (i, a, b) -> (byte)(a ^ b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, byte.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, ByteVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<ByteVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, ByteVector.class);
+    ImplCache<Associative, ReductionOperation<ByteVector, VectorMask<Byte>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, ByteVector.class);
+
+    private static ReductionOperation<ByteVector, VectorMask<Byte>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((byte)0, m, (i, a, b) -> (byte)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((byte)1, m, (i, a, b) -> (byte)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (byte) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (byte) Math.max(a, b)));
+            case VECTOR_OP_AND: return (v, m) ->
+                    toBits(v.rOp((byte)-1, m, (i, a, b) -> (byte)(a & b)));
+            case VECTOR_OP_OR: return (v, m) ->
+                    toBits(v.rOp((byte)0, m, (i, a, b) -> (byte)(a | b)));
+            case VECTOR_OP_XOR: return (v, m) ->
+                    toBits(v.rOp((byte)0, m, (i, a, b) -> (byte)(a ^ b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double128Vector.java
index 10084ccf83f..e626b706808 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double128Vector.java
@@ -236,8 +236,8 @@ Double128Vector tOp(Vector<Double> v1, Vector<Double> v2,
 
     @ForceInline
     final @Override
-    double rOp(double v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    double rOp(double v, VectorMask<Double> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final double reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final double reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Double> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Double128Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Double> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Double128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double256Vector.java
index 3d511ac336d..2d4514ebc98 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double256Vector.java
@@ -236,8 +236,8 @@ Double256Vector tOp(Vector<Double> v1, Vector<Double> v2,
 
     @ForceInline
     final @Override
-    double rOp(double v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    double rOp(double v, VectorMask<Double> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final double reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final double reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Double> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Double256Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Double> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Double256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double512Vector.java
index 1d452e88dd0..aa14a66a467 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double512Vector.java
@@ -236,8 +236,8 @@ Double512Vector tOp(Vector<Double> v1, Vector<Double> v2,
 
     @ForceInline
     final @Override
-    double rOp(double v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    double rOp(double v, VectorMask<Double> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final double reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final double reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Double> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Double512Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Double> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Double512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double64Vector.java
index ba40a8145aa..8482ddb8407 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Double64Vector.java
@@ -236,8 +236,8 @@ Double64Vector tOp(Vector<Double> v1, Vector<Double> v2,
 
     @ForceInline
     final @Override
-    double rOp(double v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    double rOp(double v, VectorMask<Double> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final double reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final double reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Double> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Double64Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Double> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Double64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleMaxVector.java
index 46c95507b29..ccba22fe2b2 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleMaxVector.java
@@ -236,8 +236,8 @@ DoubleMaxVector tOp(Vector<Double> v1, Vector<Double> v2,
 
     @ForceInline
     final @Override
-    double rOp(double v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    double rOp(double v, VectorMask<Double> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final double reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final double reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Double> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, DoubleMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Double> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, DoubleMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleVector.java
index 531e8125020..a4f3ac5a8b9 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/DoubleVector.java
@@ -287,7 +287,22 @@ DoubleVector tOpTemplate(Vector<Double> o1,
 
     /*package-private*/
     abstract
-    double rOp(double v, FBinOp f);
+    double rOp(double v, VectorMask<Double> m, FBinOp f);
+
+    @ForceInline
+    final
+    double rOpTemplate(double v, VectorMask<Double> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        double[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Double>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     double rOpTemplate(double v, FBinOp f) {
@@ -2389,9 +2404,18 @@ public abstract double reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     double reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Double>> maskClass,
                                VectorMask<Double> m) {
-        DoubleVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            DoubleVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, double.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, DoubleVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2406,24 +2430,28 @@ public abstract double reduceLanes(VectorOperators.Associative op,
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), double.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((double)0, (i, a, b) -> (double)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((double)1, (i, a, b) -> (double)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (double) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (double) Math.max(a, b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, double.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, DoubleVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<DoubleVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, DoubleVector.class);
+    ImplCache<Associative, ReductionOperation<DoubleVector, VectorMask<Double>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, DoubleVector.class);
+
+    private static ReductionOperation<DoubleVector, VectorMask<Double>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((double)0, m, (i, a, b) -> (double)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((double)1, m, (i, a, b) -> (double)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (double) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (double) Math.max(a, b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float128Vector.java
index 2a8d7ec17bd..3ac4f7bc269 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float128Vector.java
@@ -236,8 +236,8 @@ Float128Vector tOp(Vector<Float> v1, Vector<Float> v2,
 
     @ForceInline
     final @Override
-    float rOp(float v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    float rOp(float v, VectorMask<Float> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final float reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final float reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Float> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Float128Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Float> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Float128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float256Vector.java
index 3e6ee24fa8e..5a25e5054f1 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float256Vector.java
@@ -236,8 +236,8 @@ Float256Vector tOp(Vector<Float> v1, Vector<Float> v2,
 
     @ForceInline
     final @Override
-    float rOp(float v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    float rOp(float v, VectorMask<Float> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final float reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final float reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Float> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Float256Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Float> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Float256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float512Vector.java
index 54dbfabeb55..b4fea785dca 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float512Vector.java
@@ -236,8 +236,8 @@ Float512Vector tOp(Vector<Float> v1, Vector<Float> v2,
 
     @ForceInline
     final @Override
-    float rOp(float v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    float rOp(float v, VectorMask<Float> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final float reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final float reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Float> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Float512Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Float> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Float512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float64Vector.java
index 08dc6acb341..bf1bf909ced 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Float64Vector.java
@@ -236,8 +236,8 @@ Float64Vector tOp(Vector<Float> v1, Vector<Float> v2,
 
     @ForceInline
     final @Override
-    float rOp(float v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    float rOp(float v, VectorMask<Float> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final float reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final float reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Float> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Float64Mask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Float> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Float64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatMaxVector.java
index 2545228fd1d..8918d6182e5 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatMaxVector.java
@@ -236,8 +236,8 @@ FloatMaxVector tOp(Vector<Float> v1, Vector<Float> v2,
 
     @ForceInline
     final @Override
-    float rOp(float v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    float rOp(float v, VectorMask<Float> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -328,7 +328,7 @@ public final float reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final float reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Float> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, FloatMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -341,7 +341,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Float> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, FloatMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatVector.java
index 0da08805a05..474cf7c5d84 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/FloatVector.java
@@ -287,7 +287,22 @@ FloatVector tOpTemplate(Vector<Float> o1,
 
     /*package-private*/
     abstract
-    float rOp(float v, FBinOp f);
+    float rOp(float v, VectorMask<Float> m, FBinOp f);
+
+    @ForceInline
+    final
+    float rOpTemplate(float v, VectorMask<Float> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        float[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Float>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     float rOpTemplate(float v, FBinOp f) {
@@ -2409,9 +2424,18 @@ public abstract float reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     float reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Float>> maskClass,
                                VectorMask<Float> m) {
-        FloatVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            FloatVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, float.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, FloatVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2426,24 +2450,28 @@ float reduceLanesTemplate(VectorOperators.Associative op) {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), float.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((float)0, (i, a, b) -> (float)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((float)1, (i, a, b) -> (float)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (float) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (float) Math.max(a, b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, float.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, FloatVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<FloatVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, FloatVector.class);
+    ImplCache<Associative, ReductionOperation<FloatVector, VectorMask<Float>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, FloatVector.class);
+
+    private static ReductionOperation<FloatVector, VectorMask<Float>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((float)0, m, (i, a, b) -> (float)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((float)1, m, (i, a, b) -> (float)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (float) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (float) Math.max(a, b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int128Vector.java
index 32a92f0813e..3ae78e2594d 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int128Vector.java
@@ -236,8 +236,8 @@ Int128Vector tOp(Vector<Integer> v1, Vector<Integer> v2,
 
     @ForceInline
     final @Override
-    int rOp(int v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    int rOp(int v, VectorMask<Integer> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final int reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final int reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Integer> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Int128Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Integer> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Int128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int256Vector.java
index 59e07b32913..2f7a4088d5e 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int256Vector.java
@@ -236,8 +236,8 @@ Int256Vector tOp(Vector<Integer> v1, Vector<Integer> v2,
 
     @ForceInline
     final @Override
-    int rOp(int v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    int rOp(int v, VectorMask<Integer> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final int reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final int reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Integer> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Int256Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Integer> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Int256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int512Vector.java
index 91566126983..4ef56453ec8 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int512Vector.java
@@ -236,8 +236,8 @@ Int512Vector tOp(Vector<Integer> v1, Vector<Integer> v2,
 
     @ForceInline
     final @Override
-    int rOp(int v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    int rOp(int v, VectorMask<Integer> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final int reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final int reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Integer> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Int512Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Integer> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Int512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int64Vector.java
index e1dc0c69e87..51549a09845 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Int64Vector.java
@@ -236,8 +236,8 @@ Int64Vector tOp(Vector<Integer> v1, Vector<Integer> v2,
 
     @ForceInline
     final @Override
-    int rOp(int v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    int rOp(int v, VectorMask<Integer> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final int reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final int reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Integer> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Int64Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Integer> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Int64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntMaxVector.java
index aba1ef6af55..4a33df9b0d7 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntMaxVector.java
@@ -236,8 +236,8 @@ IntMaxVector tOp(Vector<Integer> v1, Vector<Integer> v2,
 
     @ForceInline
     final @Override
-    int rOp(int v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    int rOp(int v, VectorMask<Integer> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final int reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final int reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Integer> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, IntMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Integer> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, IntMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntVector.java
index 156dafcc0cf..d2cc8e613d7 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/IntVector.java
@@ -287,7 +287,22 @@ IntVector tOpTemplate(Vector<Integer> o1,
 
     /*package-private*/
     abstract
-    int rOp(int v, FBinOp f);
+    int rOp(int v, VectorMask<Integer> m, FBinOp f);
+
+    @ForceInline
+    final
+    int rOpTemplate(int v, VectorMask<Integer> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        int[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Integer>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     int rOpTemplate(int v, FBinOp f) {
@@ -2531,9 +2546,18 @@ public abstract int reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     int reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Integer>> maskClass,
                                VectorMask<Integer> m) {
-        IntVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            IntVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, int.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, IntVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2548,30 +2572,34 @@ int reduceLanesTemplate(VectorOperators.Associative op) {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), int.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((int)0, (i, a, b) -> (int)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((int)1, (i, a, b) -> (int)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (int) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (int) Math.max(a, b)));
-              case VECTOR_OP_AND: return v ->
-                      toBits(v.rOp((int)-1, (i, a, b) -> (int)(a & b)));
-              case VECTOR_OP_OR: return v ->
-                      toBits(v.rOp((int)0, (i, a, b) -> (int)(a | b)));
-              case VECTOR_OP_XOR: return v ->
-                      toBits(v.rOp((int)0, (i, a, b) -> (int)(a ^ b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, int.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, IntVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<IntVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, IntVector.class);
+    ImplCache<Associative, ReductionOperation<IntVector, VectorMask<Integer>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, IntVector.class);
+
+    private static ReductionOperation<IntVector, VectorMask<Integer>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((int)0, m, (i, a, b) -> (int)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((int)1, m, (i, a, b) -> (int)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (int) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (int) Math.max(a, b)));
+            case VECTOR_OP_AND: return (v, m) ->
+                    toBits(v.rOp((int)-1, m, (i, a, b) -> (int)(a & b)));
+            case VECTOR_OP_OR: return (v, m) ->
+                    toBits(v.rOp((int)0, m, (i, a, b) -> (int)(a | b)));
+            case VECTOR_OP_XOR: return (v, m) ->
+                    toBits(v.rOp((int)0, m, (i, a, b) -> (int)(a ^ b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long128Vector.java
index e65a8ffbba2..12ad9211668 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long128Vector.java
@@ -231,8 +231,8 @@ Long128Vector tOp(Vector<Long> v1, Vector<Long> v2,
 
     @ForceInline
     final @Override
-    long rOp(long v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    long rOp(long v, VectorMask<Long> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -329,7 +329,7 @@ public final long reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Long> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Long128Mask.class, m);  // specialized
     }
 
     @Override
@@ -342,7 +342,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Long> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Long128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long256Vector.java
index 72c4fc31336..f95bab8fcac 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long256Vector.java
@@ -231,8 +231,8 @@ Long256Vector tOp(Vector<Long> v1, Vector<Long> v2,
 
     @ForceInline
     final @Override
-    long rOp(long v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    long rOp(long v, VectorMask<Long> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -329,7 +329,7 @@ public final long reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Long> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Long256Mask.class, m);  // specialized
     }
 
     @Override
@@ -342,7 +342,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Long> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Long256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long512Vector.java
index e34f91063b7..2a047181fa5 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long512Vector.java
@@ -231,8 +231,8 @@ Long512Vector tOp(Vector<Long> v1, Vector<Long> v2,
 
     @ForceInline
     final @Override
-    long rOp(long v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    long rOp(long v, VectorMask<Long> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -329,7 +329,7 @@ public final long reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Long> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Long512Mask.class, m);  // specialized
     }
 
     @Override
@@ -342,7 +342,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Long> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Long512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long64Vector.java
index 752ed385ce5..52f13bc91c9 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Long64Vector.java
@@ -231,8 +231,8 @@ Long64Vector tOp(Vector<Long> v1, Vector<Long> v2,
 
     @ForceInline
     final @Override
-    long rOp(long v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    long rOp(long v, VectorMask<Long> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -329,7 +329,7 @@ public final long reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Long> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Long64Mask.class, m);  // specialized
     }
 
     @Override
@@ -342,7 +342,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Long> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Long64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongMaxVector.java
index e1641d27306..651428b3b27 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongMaxVector.java
@@ -231,8 +231,8 @@ LongMaxVector tOp(Vector<Long> v1, Vector<Long> v2,
 
     @ForceInline
     final @Override
-    long rOp(long v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    long rOp(long v, VectorMask<Long> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -329,7 +329,7 @@ public final long reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Long> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, LongMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -342,7 +342,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Long> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, LongMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongVector.java
index c44760a4a53..56a7a58cae9 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/LongVector.java
@@ -287,7 +287,22 @@ LongVector tOpTemplate(Vector<Long> o1,
 
     /*package-private*/
     abstract
-    long rOp(long v, FBinOp f);
+    long rOp(long v, VectorMask<Long> m, FBinOp f);
+
+    @ForceInline
+    final
+    long rOpTemplate(long v, VectorMask<Long> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        long[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Long>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     long rOpTemplate(long v, FBinOp f) {
@@ -2397,9 +2412,18 @@ public abstract long reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     long reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Long>> maskClass,
                                VectorMask<Long> m) {
-        LongVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            LongVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, long.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, LongVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2414,30 +2438,34 @@ long reduceLanesTemplate(VectorOperators.Associative op) {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), long.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((long)0, (i, a, b) -> (long)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((long)1, (i, a, b) -> (long)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (long) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (long) Math.max(a, b)));
-              case VECTOR_OP_AND: return v ->
-                      toBits(v.rOp((long)-1, (i, a, b) -> (long)(a & b)));
-              case VECTOR_OP_OR: return v ->
-                      toBits(v.rOp((long)0, (i, a, b) -> (long)(a | b)));
-              case VECTOR_OP_XOR: return v ->
-                      toBits(v.rOp((long)0, (i, a, b) -> (long)(a ^ b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, long.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, LongVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<LongVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, LongVector.class);
+    ImplCache<Associative, ReductionOperation<LongVector, VectorMask<Long>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, LongVector.class);
+
+    private static ReductionOperation<LongVector, VectorMask<Long>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((long)0, m, (i, a, b) -> (long)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((long)1, m, (i, a, b) -> (long)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (long) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (long) Math.max(a, b)));
+            case VECTOR_OP_AND: return (v, m) ->
+                    toBits(v.rOp((long)-1, m, (i, a, b) -> (long)(a & b)));
+            case VECTOR_OP_OR: return (v, m) ->
+                    toBits(v.rOp((long)0, m, (i, a, b) -> (long)(a | b)));
+            case VECTOR_OP_XOR: return (v, m) ->
+                    toBits(v.rOp((long)0, m, (i, a, b) -> (long)(a ^ b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short128Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short128Vector.java
index 77277538d21..7cceece588c 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short128Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short128Vector.java
@@ -236,8 +236,8 @@ Short128Vector tOp(Vector<Short> v1, Vector<Short> v2,
 
     @ForceInline
     final @Override
-    short rOp(short v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    short rOp(short v, VectorMask<Short> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final short reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final short reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Short> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Short128Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Short> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Short128Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short256Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short256Vector.java
index 643c2da2d51..ade7ab8bc79 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short256Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short256Vector.java
@@ -236,8 +236,8 @@ Short256Vector tOp(Vector<Short> v1, Vector<Short> v2,
 
     @ForceInline
     final @Override
-    short rOp(short v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    short rOp(short v, VectorMask<Short> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final short reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final short reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Short> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Short256Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Short> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Short256Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short512Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short512Vector.java
index b81563aa0e9..6b4f4a11796 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short512Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short512Vector.java
@@ -236,8 +236,8 @@ Short512Vector tOp(Vector<Short> v1, Vector<Short> v2,
 
     @ForceInline
     final @Override
-    short rOp(short v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    short rOp(short v, VectorMask<Short> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final short reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final short reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Short> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Short512Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Short> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Short512Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short64Vector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short64Vector.java
index f35858f0ce8..2b47fbfeacc 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short64Vector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/Short64Vector.java
@@ -236,8 +236,8 @@ Short64Vector tOp(Vector<Short> v1, Vector<Short> v2,
 
     @ForceInline
     final @Override
-    short rOp(short v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    short rOp(short v, VectorMask<Short> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final short reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final short reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Short> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, Short64Mask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Short> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, Short64Mask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortMaxVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortMaxVector.java
index 8c92a66edbe..b204fe48711 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortMaxVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortMaxVector.java
@@ -236,8 +236,8 @@ ShortMaxVector tOp(Vector<Short> v1, Vector<Short> v2,
 
     @ForceInline
     final @Override
-    short rOp(short v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    short rOp(short v, VectorMask<Short> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -334,7 +334,7 @@ public final short reduceLanes(VectorOperators.Associative op) {
     @ForceInline
     public final short reduceLanes(VectorOperators.Associative op,
                                     VectorMask<Short> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, ShortMaxMask.class, m);  // specialized
     }
 
     @Override
@@ -347,7 +347,7 @@ public final long reduceLanesToLong(VectorOperators.Associative op) {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<Short> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, ShortMaxMask.class, m);  // specialized
     }
 
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortVector.java b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortVector.java
index 03867ac7320..681d160c2af 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortVector.java
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/ShortVector.java
@@ -287,7 +287,22 @@ ShortVector tOpTemplate(Vector<Short> o1,
 
     /*package-private*/
     abstract
-    short rOp(short v, FBinOp f);
+    short rOp(short v, VectorMask<Short> m, FBinOp f);
+
+    @ForceInline
+    final
+    short rOpTemplate(short v, VectorMask<Short> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        short[] vec = vec();
+        boolean[] mbits = ((AbstractMask<Short>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     short rOpTemplate(short v, FBinOp f) {
@@ -2532,9 +2547,18 @@ public abstract short reduceLanes(VectorOperators.Associative op,
     @ForceInline
     final
     short reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<Short>> maskClass,
                                VectorMask<Short> m) {
-        ShortVector v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            ShortVector v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, short.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, ShortVector::reductionOperations)));
     }
 
     /*package-private*/
@@ -2549,30 +2573,34 @@ short reduceLanesTemplate(VectorOperators.Associative op) {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), short.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp((short)0, (i, a, b) -> (short)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp((short)1, (i, a, b) -> (short)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> (short) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> (short) Math.max(a, b)));
-              case VECTOR_OP_AND: return v ->
-                      toBits(v.rOp((short)-1, (i, a, b) -> (short)(a & b)));
-              case VECTOR_OP_OR: return v ->
-                      toBits(v.rOp((short)0, (i, a, b) -> (short)(a | b)));
-              case VECTOR_OP_XOR: return v ->
-                      toBits(v.rOp((short)0, (i, a, b) -> (short)(a ^ b)));
-              default: return null;
-              }})));
+            opc, getClass(), null, short.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, ShortVector::reductionOperations)));
     }
+
     private static final
-    ImplCache<Associative,Function<ShortVector,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, ShortVector.class);
+    ImplCache<Associative, ReductionOperation<ShortVector, VectorMask<Short>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, ShortVector.class);
+
+    private static ReductionOperation<ShortVector, VectorMask<Short>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp((short)0, m, (i, a, b) -> (short)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp((short)1, m, (i, a, b) -> (short)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> (short) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> (short) Math.max(a, b)));
+            case VECTOR_OP_AND: return (v, m) ->
+                    toBits(v.rOp((short)-1, m, (i, a, b) -> (short)(a & b)));
+            case VECTOR_OP_OR: return (v, m) ->
+                    toBits(v.rOp((short)0, m, (i, a, b) -> (short)(a | b)));
+            case VECTOR_OP_XOR: return (v, m) ->
+                    toBits(v.rOp((short)0, m, (i, a, b) -> (short)(a ^ b)));
+            default: return null;
+        }
+    }
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-Vector.java.template b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-Vector.java.template
index 306c1eb8cb7..0b1bfe82de5 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-Vector.java.template
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-Vector.java.template
@@ -291,7 +291,22 @@ public abstract class $abstractvectortype$ extends AbstractVector<$Boxtype$> {
 
     /*package-private*/
     abstract
-    $type$ rOp($type$ v, FBinOp f);
+    $type$ rOp($type$ v, VectorMask<$Boxtype$> m, FBinOp f);
+
+    @ForceInline
+    final
+    $type$ rOpTemplate($type$ v, VectorMask<$Boxtype$> m, FBinOp f) {
+        if (m == null) {
+            return rOpTemplate(v, f);
+        }
+        $type$[] vec = vec();
+        boolean[] mbits = ((AbstractMask<$Boxtype$>)m).getBits();
+        for (int i = 0; i < vec.length; i++) {
+            v = mbits[i] ? f.apply(i, v, vec[i]) : v;
+        }
+        return v;
+    }
+
     @ForceInline
     final
     $type$ rOpTemplate($type$ v, FBinOp f) {
@@ -2948,9 +2963,18 @@ public abstract class $abstractvectortype$ extends AbstractVector<$Boxtype$> {
     @ForceInline
     final
     $type$ reduceLanesTemplate(VectorOperators.Associative op,
+                               Class<? extends VectorMask<$Boxtype$>> maskClass,
                                VectorMask<$Boxtype$> m) {
-        $abstractvectortype$ v = reduceIdentityVector(op).blend(this, m);
-        return v.reduceLanesTemplate(op);
+        m.check(maskClass, this);
+        if (op == FIRST_NONZERO) {
+            $abstractvectortype$ v = reduceIdentityVector(op).blend(this, m);
+            return v.reduceLanesTemplate(op);
+        }
+        int opc = opCode(op);
+        return fromBits(VectorSupport.reductionCoerced(
+            opc, getClass(), maskClass, $type$.class, length(),
+            this, m,
+            REDUCE_IMPL.find(op, opc, $abstractvectortype$::reductionOperations)));
     }
 
     /*package-private*/
@@ -2965,32 +2989,36 @@ public abstract class $abstractvectortype$ extends AbstractVector<$Boxtype$> {
         }
         int opc = opCode(op);
         return fromBits(VectorSupport.reductionCoerced(
-            opc, getClass(), $type$.class, length(),
-            this,
-            REDUCE_IMPL.find(op, opc, (opc_) -> {
-              switch (opc_) {
-              case VECTOR_OP_ADD: return v ->
-                      toBits(v.rOp(($type$)0, (i, a, b) -> ($type$)(a + b)));
-              case VECTOR_OP_MUL: return v ->
-                      toBits(v.rOp(($type$)1, (i, a, b) -> ($type$)(a * b)));
-              case VECTOR_OP_MIN: return v ->
-                      toBits(v.rOp(MAX_OR_INF, (i, a, b) -> ($type$) Math.min(a, b)));
-              case VECTOR_OP_MAX: return v ->
-                      toBits(v.rOp(MIN_OR_INF, (i, a, b) -> ($type$) Math.max(a, b)));
+            opc, getClass(), null, $type$.class, length(),
+            this, null,
+            REDUCE_IMPL.find(op, opc, $abstractvectortype$::reductionOperations)));
+    }
+
+    private static final
+    ImplCache<Associative, ReductionOperation<$abstractvectortype$, VectorMask<$Boxtype$>>>
+        REDUCE_IMPL = new ImplCache<>(Associative.class, $Type$Vector.class);
+
+    private static ReductionOperation<$abstractvectortype$, VectorMask<$Boxtype$>> reductionOperations(int opc_) {
+        switch (opc_) {
+            case VECTOR_OP_ADD: return (v, m) ->
+                    toBits(v.rOp(($type$)0, m, (i, a, b) -> ($type$)(a + b)));
+            case VECTOR_OP_MUL: return (v, m) ->
+                    toBits(v.rOp(($type$)1, m, (i, a, b) -> ($type$)(a * b)));
+            case VECTOR_OP_MIN: return (v, m) ->
+                    toBits(v.rOp(MAX_OR_INF, m, (i, a, b) -> ($type$) Math.min(a, b)));
+            case VECTOR_OP_MAX: return (v, m) ->
+                    toBits(v.rOp(MIN_OR_INF, m, (i, a, b) -> ($type$) Math.max(a, b)));
 #if[BITWISE]
-              case VECTOR_OP_AND: return v ->
-                      toBits(v.rOp(($type$)-1, (i, a, b) -> ($type$)(a & b)));
-              case VECTOR_OP_OR: return v ->
-                      toBits(v.rOp(($type$)0, (i, a, b) -> ($type$)(a | b)));
-              case VECTOR_OP_XOR: return v ->
-                      toBits(v.rOp(($type$)0, (i, a, b) -> ($type$)(a ^ b)));
+            case VECTOR_OP_AND: return (v, m) ->
+                    toBits(v.rOp(($type$)-1, m, (i, a, b) -> ($type$)(a & b)));
+            case VECTOR_OP_OR: return (v, m) ->
+                    toBits(v.rOp(($type$)0, m, (i, a, b) -> ($type$)(a | b)));
+            case VECTOR_OP_XOR: return (v, m) ->
+                    toBits(v.rOp(($type$)0, m, (i, a, b) -> ($type$)(a ^ b)));
 #end[BITWISE]
-              default: return null;
-              }})));
+            default: return null;
+        }
     }
-    private static final
-    ImplCache<Associative,Function<$abstractvectortype$,Long>> REDUCE_IMPL
-        = new ImplCache<>(Associative.class, $Type$Vector.class);
 
     private
     @ForceInline
diff --git a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-VectorBits.java.template b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-VectorBits.java.template
index 54b50fb7cbd..a5fe360f21c 100644
--- a/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-VectorBits.java.template
+++ b/src/jdk.incubator.vector/share/classes/jdk/incubator/vector/X-VectorBits.java.template
@@ -238,8 +238,8 @@ final class $vectortype$ extends $abstractvectortype$ {
 
     @ForceInline
     final @Override
-    $type$ rOp($type$ v, FBinOp f) {
-        return super.rOpTemplate(v, f);  // specialize
+    $type$ rOp($type$ v, VectorMask<$Boxtype$> m, FBinOp f) {
+        return super.rOpTemplate(v, m, f);  // specialize
     }
 
     @Override
@@ -338,7 +338,7 @@ final class $vectortype$ extends $abstractvectortype$ {
     @ForceInline
     public final $type$ reduceLanes(VectorOperators.Associative op,
                                     VectorMask<$Boxtype$> m) {
-        return super.reduceLanesTemplate(op, m);  // specialized
+        return super.reduceLanesTemplate(op, $masktype$.class, m);  // specialized
     }
 
     @Override
@@ -351,7 +351,7 @@ final class $vectortype$ extends $abstractvectortype$ {
     @ForceInline
     public final long reduceLanesToLong(VectorOperators.Associative op,
                                         VectorMask<$Boxtype$> m) {
-        return (long) super.reduceLanesTemplate(op, m);  // specialized
+        return (long) super.reduceLanesTemplate(op, $masktype$.class, m);  // specialized
     }
 
     @ForceInline