Make redux_vec_unroller more flexible regarding packet-type

1 files changed, 23 insertions, 25 deletions

diff --git a/Eigen/src/Core/Redux.h b/Eigen/src/Core/Redux.h
index 0aee855df..720b6030c 100644
--- a/Eigen/src/Core/Redux.h
+++ b/Eigen/src/Core/Redux.h
@@ -145,38 +145,36 @@ struct redux_novec_unroller<Func, Evaluator, Start, 0>
 template<typename Func, typename Evaluator, int Start, int Length>
 struct redux_vec_unroller
 {
-  enum {
-    PacketSize = redux_traits<Func, Evaluator>::PacketSize,
-    HalfLength = Length/2
-  };
-
-  typedef typename Evaluator::Scalar Scalar;
-  typedef typename redux_traits<Func, Evaluator>::PacketType PacketScalar;
-
-  static EIGEN_STRONG_INLINE PacketScalar run(const Evaluator &eval, const Func& func)
+  template<typename PacketType>
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE PacketType run(const Evaluator &eval, const Func& func)
   {
+    enum {
+      PacketSize = unpacket_traits<PacketType>::size,
+      HalfLength = Length/2
+    };
+
     return func.packetOp(
-            redux_vec_unroller<Func, Evaluator, Start, HalfLength>::run(eval,func),
-            redux_vec_unroller<Func, Evaluator, Start+HalfLength, Length-HalfLength>::run(eval,func) );
+            redux_vec_unroller<Func, Evaluator, Start, HalfLength>::template run<PacketType>(eval,func),
+            redux_vec_unroller<Func, Evaluator, Start+HalfLength, Length-HalfLength>::template run<PacketType>(eval,func) );
   }
 };
 
 template<typename Func, typename Evaluator, int Start>
 struct redux_vec_unroller<Func, Evaluator, Start, 1>
 {
-  enum {
-    index = Start * redux_traits<Func, Evaluator>::PacketSize,
-    outer = index / int(Evaluator::InnerSizeAtCompileTime),
-    inner = index % int(Evaluator::InnerSizeAtCompileTime),
-    alignment = Evaluator::Alignment
-  };
-
-  typedef typename Evaluator::Scalar Scalar;
-  typedef typename redux_traits<Func, Evaluator>::PacketType PacketScalar;
-
-  static EIGEN_STRONG_INLINE PacketScalar run(const Evaluator &eval, const Func&)
+  template<typename PacketType>
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE PacketType run(const Evaluator &eval, const Func&)
   {
-    return eval.template packetByOuterInner<alignment,PacketScalar>(outer, inner);
+    enum {
+      PacketSize = unpacket_traits<PacketType>::size,
+      index = Start * PacketSize,
+      outer = index / int(Evaluator::InnerSizeAtCompileTime),
+      inner = index % int(Evaluator::InnerSizeAtCompileTime),
+      alignment = Evaluator::Alignment
+    };
+    return eval.template packetByOuterInner<alignment,PacketType>(outer, inner);
   }
 };
 
@@ -329,7 +327,7 @@ struct redux_impl<Func, Evaluator, LinearVectorizedTraversal, CompleteUnrolling>
 {
   typedef typename Evaluator::Scalar Scalar;
 
-  typedef typename redux_traits<Func, Evaluator>::PacketType PacketScalar;
+  typedef typename redux_traits<Func, Evaluator>::PacketType PacketType;
   enum {
     PacketSize = redux_traits<Func, Evaluator>::PacketSize,
     Size = Evaluator::SizeAtCompileTime,
@@ -343,7 +341,7 @@ struct redux_impl<Func, Evaluator, LinearVectorizedTraversal, CompleteUnrolling>
     EIGEN_ONLY_USED_FOR_DEBUG(xpr)
     eigen_assert(xpr.rows()>0 && xpr.cols()>0 && "you are using an empty matrix");
     if (VectorizedSize > 0) {
-      Scalar res = func.predux(redux_vec_unroller<Func, Evaluator, 0, Size / PacketSize>::run(eval,func));
+      Scalar res = func.predux(redux_vec_unroller<Func, Evaluator, 0, Size / PacketSize>::template run<PacketType>(eval,func));
       if (VectorizedSize != Size)
         res = func(res,redux_novec_unroller<Func, Evaluator, VectorizedSize, Size-VectorizedSize>::run(eval,func));
       return res;