Add support for NVCC5: most of the Core and part of LU are callable from CUDA code.

Still a lot to do.

1 files changed, 16 insertions, 0 deletions

diff --git a/Eigen/src/Core/Assign.h b/Eigen/src/Core/Assign.h
index cd29a88f0..dc9b55fa4 100644
--- a/Eigen/src/Core/Assign.h
+++ b/Eigen/src/Core/Assign.h
@@ -139,6 +139,7 @@ struct assign_DefaultTraversal_CompleteUnrolling
     inner = Index % Derived1::InnerSizeAtCompileTime
   };
 
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src)
   {
     dst.copyCoeffByOuterInner(outer, inner, src);
@@ -149,12 +150,14 @@ struct assign_DefaultTraversal_CompleteUnrolling
 template<typename Derived1, typename Derived2, int Stop>
 struct assign_DefaultTraversal_CompleteUnrolling<Derived1, Derived2, Stop, Stop>
 {
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &, const Derived2 &) {}
 };
 
 template<typename Derived1, typename Derived2, int Index, int Stop>
 struct assign_DefaultTraversal_InnerUnrolling
 {
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src, int outer)
   {
     dst.copyCoeffByOuterInner(outer, Index, src);
@@ -165,6 +168,7 @@ struct assign_DefaultTraversal_InnerUnrolling
 template<typename Derived1, typename Derived2, int Stop>
 struct assign_DefaultTraversal_InnerUnrolling<Derived1, Derived2, Stop, Stop>
 {
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &, const Derived2 &, int) {}
 };
 
@@ -175,6 +179,7 @@ struct assign_DefaultTraversal_InnerUnrolling<Derived1, Derived2, Stop, Stop>
 template<typename Derived1, typename Derived2, int Index, int Stop>
 struct assign_LinearTraversal_CompleteUnrolling
 {
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src)
   {
     dst.copyCoeff(Index, src);
@@ -185,6 +190,7 @@ struct assign_LinearTraversal_CompleteUnrolling
 template<typename Derived1, typename Derived2, int Stop>
 struct assign_LinearTraversal_CompleteUnrolling<Derived1, Derived2, Stop, Stop>
 {
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &, const Derived2 &) {}
 };
 
@@ -249,6 +255,7 @@ struct assign_impl;
 template<typename Derived1, typename Derived2, int Unrolling, int Version>
 struct assign_impl<Derived1, Derived2, InvalidTraversal, Unrolling, Version>
 {
+  EIGEN_DEVICE_FUNC
   static inline void run(Derived1 &, const Derived2 &) { }
 };
 
@@ -256,6 +263,7 @@ template<typename Derived1, typename Derived2, int Version>
 struct assign_impl<Derived1, Derived2, DefaultTraversal, NoUnrolling, Version>
 {
   typedef typename Derived1::Index Index;
+  EIGEN_DEVICE_FUNC 
   static inline void run(Derived1 &dst, const Derived2 &src)
   {
     const Index innerSize = dst.innerSize();
@@ -269,6 +277,7 @@ struct assign_impl<Derived1, Derived2, DefaultTraversal, NoUnrolling, Version>
 template<typename Derived1, typename Derived2, int Version>
 struct assign_impl<Derived1, Derived2, DefaultTraversal, CompleteUnrolling, Version>
 {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src)
   {
     assign_DefaultTraversal_CompleteUnrolling<Derived1, Derived2, 0, Derived1::SizeAtCompileTime>
@@ -280,6 +289,7 @@ template<typename Derived1, typename Derived2, int Version>
 struct assign_impl<Derived1, Derived2, DefaultTraversal, InnerUnrolling, Version>
 {
   typedef typename Derived1::Index Index;
+  EIGEN_DEVICE_FUNC 
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src)
   {
     const Index outerSize = dst.outerSize();
@@ -297,6 +307,7 @@ template<typename Derived1, typename Derived2, int Version>
 struct assign_impl<Derived1, Derived2, LinearTraversal, NoUnrolling, Version>
 {
   typedef typename Derived1::Index Index;
+  EIGEN_DEVICE_FUNC
   static inline void run(Derived1 &dst, const Derived2 &src)
   {
     const Index size = dst.size();
@@ -308,6 +319,7 @@ struct assign_impl<Derived1, Derived2, LinearTraversal, NoUnrolling, Version>
 template<typename Derived1, typename Derived2, int Version>
 struct assign_impl<Derived1, Derived2, LinearTraversal, CompleteUnrolling, Version>
 {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE void run(Derived1 &dst, const Derived2 &src)
   {
     assign_LinearTraversal_CompleteUnrolling<Derived1, Derived2, 0, Derived1::SizeAtCompileTime>
@@ -519,18 +531,22 @@ struct assign_selector;
 
 template<typename Derived, typename OtherDerived>
 struct assign_selector<Derived,OtherDerived,false,false> {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE Derived& run(Derived& dst, const OtherDerived& other) { return dst.lazyAssign(other.derived()); }
 };
 template<typename Derived, typename OtherDerived>
 struct assign_selector<Derived,OtherDerived,true,false> {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE Derived& run(Derived& dst, const OtherDerived& other) { return dst.lazyAssign(other.eval()); }
 };
 template<typename Derived, typename OtherDerived>
 struct assign_selector<Derived,OtherDerived,false,true> {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE Derived& run(Derived& dst, const OtherDerived& other) { return dst.lazyAssign(other.transpose()); }
 };
 template<typename Derived, typename OtherDerived>
 struct assign_selector<Derived,OtherDerived,true,true> {
+  EIGEN_DEVICE_FUNC
   static EIGEN_STRONG_INLINE Derived& run(Derived& dst, const OtherDerived& other) { return dst.lazyAssign(other.transpose().eval()); }
 };