From b2053990d078e15017610d526445ce78f2632edf Mon Sep 17 00:00:00 2001
From: Robert Lukierski <robert@lukierski.eu>
Date: Wed, 14 Mar 2018 16:19:43 +0000
Subject: Adding EIGEN_DEVICE_FUNC to Products, especially Dense2Dense
 Assignment specializations. Otherwise causes problems with small fixed size
 matrix multiplication (call to 0x00 in call_assignment_no_alias in debug mode
 or trap in release with CUDA 9.1).

---
 Eigen/src/Core/ProductEvaluators.h | 50 +++++++++++++++++++-------------------
 1 file changed, 25 insertions(+), 25 deletions(-)

(limited to 'Eigen/src/Core/ProductEvaluators.h')

diff --git a/Eigen/src/Core/ProductEvaluators.h b/Eigen/src/Core/ProductEvaluators.h
index fb637191d..5d2737d01 100644
--- a/Eigen/src/Core/ProductEvaluators.h
+++ b/Eigen/src/Core/ProductEvaluators.h
@@ -137,7 +137,7 @@ struct Assignment<DstXprType, Product<Lhs,Rhs,Options>, internal::assign_op<Scal
   typename enable_if<(Options==DefaultProduct || Options==AliasFreeProduct)>::type>
 {
   typedef Product<Lhs,Rhs,Options> SrcXprType;
-  static EIGEN_STRONG_INLINE
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE
   void run(DstXprType &dst, const SrcXprType &src, const internal::assign_op<Scalar,Scalar> &)
   {
     Index dstRows = src.rows();
@@ -155,7 +155,7 @@ struct Assignment<DstXprType, Product<Lhs,Rhs,Options>, internal::add_assign_op<
   typename enable_if<(Options==DefaultProduct || Options==AliasFreeProduct)>::type>
 {
   typedef Product<Lhs,Rhs,Options> SrcXprType;
-  static EIGEN_STRONG_INLINE
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE
   void run(DstXprType &dst, const SrcXprType &src, const internal::add_assign_op<Scalar,Scalar> &)
   {
     eigen_assert(dst.rows() == src.rows() && dst.cols() == src.cols());
@@ -170,7 +170,7 @@ struct Assignment<DstXprType, Product<Lhs,Rhs,Options>, internal::sub_assign_op<
   typename enable_if<(Options==DefaultProduct || Options==AliasFreeProduct)>::type>
 {
   typedef Product<Lhs,Rhs,Options> SrcXprType;
-  static EIGEN_STRONG_INLINE
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE
   void run(DstXprType &dst, const SrcXprType &src, const internal::sub_assign_op<Scalar,Scalar> &)
   {
     eigen_assert(dst.rows() == src.rows() && dst.cols() == src.cols());
@@ -190,7 +190,7 @@ struct Assignment<DstXprType, CwiseBinaryOp<internal::scalar_product_op<ScalarBi
   typedef CwiseBinaryOp<internal::scalar_product_op<ScalarBis,Scalar>,
                         const CwiseNullaryOp<internal::scalar_constant_op<ScalarBis>,Plain>,
                         const Product<Lhs,Rhs,DefaultProduct> > SrcXprType;
-  static EIGEN_STRONG_INLINE
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE
   void run(DstXprType &dst, const SrcXprType &src, const AssignFunc& func)
   {
     call_assignment_no_alias(dst, (src.lhs().functor().m_other * src.rhs().lhs())*src.rhs().rhs(), func);
@@ -217,7 +217,7 @@ template<typename DstXprType, typename OtherXpr, typename ProductType, typename
 struct assignment_from_xpr_op_product
 {
   template<typename SrcXprType, typename InitialFunc>
-  static EIGEN_STRONG_INLINE
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE
   void run(DstXprType &dst, const SrcXprType &src, const InitialFunc& /*func*/)
   {
     call_assignment_no_alias(dst, src.lhs(), Func1());
@@ -246,19 +246,19 @@ template<typename Lhs, typename Rhs>
 struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,InnerProduct>
 {
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     dst.coeffRef(0,0) = (lhs.transpose().cwiseProduct(rhs)).sum();
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     dst.coeffRef(0,0) += (lhs.transpose().cwiseProduct(rhs)).sum();
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   { dst.coeffRef(0,0) -= (lhs.transpose().cwiseProduct(rhs)).sum(); }
 };
 
@@ -269,7 +269,7 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,InnerProduct>
 
 // Column major result
 template<typename Dst, typename Lhs, typename Rhs, typename Func>
-void outer_product_selector_run(Dst& dst, const Lhs &lhs, const Rhs &rhs, const Func& func, const false_type&)
+void EIGEN_DEVICE_FUNC outer_product_selector_run(Dst& dst, const Lhs &lhs, const Rhs &rhs, const Func& func, const false_type&)
 {
   evaluator<Rhs> rhsEval(rhs);
   typename nested_eval<Lhs,Rhs::SizeAtCompileTime>::type actual_lhs(lhs);
@@ -282,7 +282,7 @@ void outer_product_selector_run(Dst& dst, const Lhs &lhs, const Rhs &rhs, const
 
 // Row major result
 template<typename Dst, typename Lhs, typename Rhs, typename Func>
-void outer_product_selector_run(Dst& dst, const Lhs &lhs, const Rhs &rhs, const Func& func, const true_type&)
+void EIGEN_DEVICE_FUNC outer_product_selector_run(Dst& dst, const Lhs &lhs, const Rhs &rhs, const Func& func, const true_type&)
 {
   evaluator<Lhs> lhsEval(lhs);
   typename nested_eval<Rhs,Lhs::SizeAtCompileTime>::type actual_rhs(rhs);
@@ -300,37 +300,37 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,OuterProduct>
   typedef typename Product<Lhs,Rhs>::Scalar Scalar;
   
   // TODO it would be nice to be able to exploit our *_assign_op functors for that purpose
-  struct set  { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived()  = src; } };
-  struct add  { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() += src; } };
-  struct sub  { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() -= src; } };
+  struct set  { template<typename Dst, typename Src> EIGEN_DEVICE_FUNC void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived()  = src; } };
+  struct add  { template<typename Dst, typename Src> EIGEN_DEVICE_FUNC void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() += src; } };
+  struct sub  { template<typename Dst, typename Src> EIGEN_DEVICE_FUNC void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() -= src; } };
   struct adds {
     Scalar m_scale;
     explicit adds(const Scalar& s) : m_scale(s) {}
-    template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const {
+    template<typename Dst, typename Src> void EIGEN_DEVICE_FUNC operator()(const Dst& dst, const Src& src) const {
       dst.const_cast_derived() += m_scale * src;
     }
   };
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     internal::outer_product_selector_run(dst, lhs, rhs, set(), is_row_major<Dst>());
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     internal::outer_product_selector_run(dst, lhs, rhs, add(), is_row_major<Dst>());
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     internal::outer_product_selector_run(dst, lhs, rhs, sub(), is_row_major<Dst>());
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void scaleAndAddTo(Dst& dst, const Lhs& lhs, const Rhs& rhs, const Scalar& alpha)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void scaleAndAddTo(Dst& dst, const Lhs& lhs, const Rhs& rhs, const Scalar& alpha)
   {
     internal::outer_product_selector_run(dst, lhs, rhs, adds(alpha), is_row_major<Dst>());
   }
@@ -345,15 +345,15 @@ struct generic_product_impl_base
   typedef typename Product<Lhs,Rhs>::Scalar Scalar;
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   { dst.setZero(); scaleAndAddTo(dst, lhs, rhs, Scalar(1)); }
 
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   { scaleAndAddTo(dst,lhs, rhs, Scalar(1)); }
 
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   { scaleAndAddTo(dst, lhs, rhs, Scalar(-1)); }
   
   template<typename Dst>
@@ -373,7 +373,7 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,GemvProduct>
   typedef typename internal::remove_all<typename internal::conditional<int(Side)==OnTheRight,LhsNested,RhsNested>::type>::type MatrixType;
 
   template<typename Dest>
-  static EIGEN_STRONG_INLINE void scaleAndAddTo(Dest& dst, const Lhs& lhs, const Rhs& rhs, const Scalar& alpha)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void scaleAndAddTo(Dest& dst, const Lhs& lhs, const Rhs& rhs, const Scalar& alpha)
   {
     LhsNested actual_lhs(lhs);
     RhsNested actual_rhs(rhs);
@@ -390,7 +390,7 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,CoeffBasedProductMode>
   typedef typename Product<Lhs,Rhs>::Scalar Scalar;
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     // Same as: dst.noalias() = lhs.lazyProduct(rhs);
     // but easier on the compiler side
@@ -398,14 +398,14 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,CoeffBasedProductMode>
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void addTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     // dst.noalias() += lhs.lazyProduct(rhs);
     call_assignment_no_alias(dst, lhs.lazyProduct(rhs), internal::add_assign_op<typename Dst::Scalar,Scalar>());
   }
   
   template<typename Dst>
-  static EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
+  static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void subTo(Dst& dst, const Lhs& lhs, const Rhs& rhs)
   {
     // dst.noalias() -= lhs.lazyProduct(rhs);
     call_assignment_no_alias(dst, lhs.lazyProduct(rhs), internal::sub_assign_op<typename Dst::Scalar,Scalar>());
-- 
cgit v1.2.3