updates based on PR feedback

There are two major changes (and a few minor ones which are not listed here...see PR discussion for details)

1. Eigen::half implementations for HIP and CUDA have been merged.
This means that
- `CUDA/Half.h` and `HIP/hcc/Half.h` got merged to a new file `GPU/Half.h`
- `CUDA/PacketMathHalf.h` and `HIP/hcc/PacketMathHalf.h` got merged to a new file `GPU/PacketMathHalf.h`
- `CUDA/TypeCasting.h` and `HIP/hcc/TypeCasting.h` got merged to a new file `GPU/TypeCasting.h`

After this change the `HIP/hcc` directory only contains one file `math_constants.h`. That will go away too once that file becomes a part of the HIP install.

2. new macros EIGEN_GPUCC, EIGEN_GPU_COMPILE_PHASE and EIGEN_HAS_GPU_FP16 have been added and the code has been updated to use them where appropriate.
- `EIGEN_GPUCC` is the same as `(EIGEN_CUDACC || EIGEN_HIPCC)`
- `EIGEN_GPU_DEVICE_COMPILE` is the same as `(EIGEN_CUDA_ARCH || EIGEN_HIP_DEVICE_COMPILE)`
- `EIGEN_HAS_GPU_FP16` is the same as `(EIGEN_HAS_CUDA_FP16 or EIGEN_HAS_HIP_FP16)`

1 files changed, 4 insertions, 16 deletions

diff --git a/unsupported/Eigen/CXX11/src/util/CXX11Meta.h b/unsupported/Eigen/CXX11/src/util/CXX11Meta.h
index bb584e3f9..8de3bbcab 100644
--- a/unsupported/Eigen/CXX11/src/util/CXX11Meta.h
+++ b/unsupported/Eigen/CXX11/src/util/CXX11Meta.h
@@ -268,10 +268,7 @@ template<
   typename Reducer
 > struct reduce<Reducer>
 {
-  #if defined(EIGEN_HIPCC)
-  EIGEN_DEVICE_FUNC
-  #endif
-  constexpr static inline int run() { return Reducer::Identity; }
+  EIGEN_DEVICE_FUNC constexpr static inline int run() { return Reducer::Identity; }
 };
 
 template<
@@ -279,10 +276,7 @@ template<
   typename A
 > struct reduce<Reducer, A>
 {
-  #if defined(EIGEN_HIPCC)
-  EIGEN_DEVICE_FUNC
-  #endif
-  constexpr static inline A run(A a) { return a; }
+  EIGEN_DEVICE_FUNC constexpr static inline A run(A a) { return a; }
 };
 
 template<
@@ -291,10 +285,7 @@ template<
   typename... Ts
 > struct reduce<Reducer, A, Ts...>
 {
-  #if defined(EIGEN_HIPCC)
-  EIGEN_DEVICE_FUNC
-  #endif
-  constexpr static inline auto run(A a, Ts... ts) -> decltype(Reducer::run(a, reduce<Reducer, Ts...>::run(ts...))) {
+  EIGEN_DEVICE_FUNC constexpr static inline auto run(A a, Ts... ts) -> decltype(Reducer::run(a, reduce<Reducer, Ts...>::run(ts...))) {
     return Reducer::run(a, reduce<Reducer, Ts...>::run(ts...));
   }
 };
@@ -333,10 +324,7 @@ struct greater_equal_zero_op { template<typename A> constexpr static inline auto
 // together in front... (13.0 doesn't work with array_prod/array_reduce/... anyway, but 13.1
 // does...
 template<typename... Ts>
-#if defined(EIGEN_HIPCC)
-EIGEN_DEVICE_FUNC
-#endif
-constexpr inline decltype(reduce<product_op, Ts...>::run((*((Ts*)0))...)) arg_prod(Ts... ts)
+EIGEN_DEVICE_FUNC constexpr inline decltype(reduce<product_op, Ts...>::run((*((Ts*)0))...)) arg_prod(Ts... ts)
 {
   return reduce<product_op, Ts...>::run(ts...);
 }