[SYCL] Rebasing the SYCL support branch on top of the Einge upstream master branch.

* Unifying all loadLocalTile from lhs and rhs to an extract_block function.
* Adding get_tensor operation which was missing in TensorContractionMapper.
* Adding the -D method missing from cmake for Disable_Skinny Contraction operation.
* Wrapping all the indices in TensorScanSycl into Scan parameter struct.
* Fixing typo in Device SYCL
* Unifying load to private register for tall/skinny no shared
* Unifying load to vector tile for tensor-vector/vector-tensor operation
* Removing all the LHS/RHS class for extracting data from global
* Removing Outputfunction from TensorContractionSkinnyNoshared.
* Combining the local memory version of tall/skinny and normal tensor contraction into one kernel.
* Combining the no-local memory version of tall/skinny and normal tensor contraction into one kernel.
* Combining General Tensor-Vector and VectorTensor contraction into one kernel.
* Making double buffering optional for Tensor contraction when local memory is version is used.
* Modifying benchmark to accept custom Reduction Sizes
* Disabling AVX optimization for SYCL backend on the host to allow SSE optimization to the host
* Adding Test for SYCL
* Modifying SYCL CMake

1 files changed, 25 insertions, 27 deletions

diff --git a/unsupported/test/cxx11_tensor_shuffling_sycl.cpp b/unsupported/test/cxx11_tensor_shuffling_sycl.cpp
index 0e8cc3bd2..ca4e8b5ef 100644
--- a/unsupported/test/cxx11_tensor_shuffling_sycl.cpp
+++ b/unsupported/test/cxx11_tensor_shuffling_sycl.cpp
@@ -12,14 +12,12 @@
 // Public License v. 2.0. If a copy of the MPL was not distributed
 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
 
-
 #define EIGEN_TEST_NO_LONGDOUBLE
 #define EIGEN_TEST_NO_COMPLEX
 
 #define EIGEN_DEFAULT_DENSE_INDEX_TYPE int64_t
 #define EIGEN_USE_SYCL
 
-
 #include "main.h"
 #include <unsupported/Eigen/CXX11/Tensor>
 
@@ -29,33 +27,33 @@ using Eigen::Tensor;
 using Eigen::TensorMap;
 
 template <typename DataType, int DataLayout, typename IndexType>
-static void test_simple_shuffling_sycl(const Eigen::SyclDevice& sycl_device)
-{
+static void test_simple_shuffling_sycl(const Eigen::SyclDevice& sycl_device) {
   IndexType sizeDim1 = 2;
   IndexType sizeDim2 = 3;
   IndexType sizeDim3 = 5;
   IndexType sizeDim4 = 7;
   array<IndexType, 4> tensorRange = {{sizeDim1, sizeDim2, sizeDim3, sizeDim4}};
-  Tensor<DataType, 4, DataLayout,IndexType> tensor(tensorRange);
-  Tensor<DataType, 4, DataLayout,IndexType> no_shuffle(tensorRange);
+  Tensor<DataType, 4, DataLayout, IndexType> tensor(tensorRange);
+  Tensor<DataType, 4, DataLayout, IndexType> no_shuffle(tensorRange);
   tensor.setRandom();
 
-  const size_t buffSize =tensor.size()*sizeof(DataType);
+  const size_t buffSize = tensor.size() * sizeof(DataType);
   array<IndexType, 4> shuffles;
   shuffles[0] = 0;
   shuffles[1] = 1;
   shuffles[2] = 2;
   shuffles[3] = 3;
-  DataType* gpu_data1  = static_cast<DataType*>(sycl_device.allocate(buffSize));
-  DataType* gpu_data2  = static_cast<DataType*>(sycl_device.allocate(buffSize));
-
+  DataType* gpu_data1 = static_cast<DataType*>(sycl_device.allocate(buffSize));
+  DataType* gpu_data2 = static_cast<DataType*>(sycl_device.allocate(buffSize));
 
-  TensorMap<Tensor<DataType, 4, DataLayout,IndexType>> gpu1(gpu_data1, tensorRange);
-  TensorMap<Tensor<DataType, 4, DataLayout,IndexType>> gpu2(gpu_data2, tensorRange);
+  TensorMap<Tensor<DataType, 4, DataLayout, IndexType>> gpu1(gpu_data1,
+                                                             tensorRange);
+  TensorMap<Tensor<DataType, 4, DataLayout, IndexType>> gpu2(gpu_data2,
+                                                             tensorRange);
 
   sycl_device.memcpyHostToDevice(gpu_data1, tensor.data(), buffSize);
 
-  gpu2.device(sycl_device)=gpu1.shuffle(shuffles);
+  gpu2.device(sycl_device) = gpu1.shuffle(shuffles);
   sycl_device.memcpyDeviceToHost(no_shuffle.data(), gpu_data2, buffSize);
   sycl_device.synchronize();
 
@@ -68,7 +66,7 @@ static void test_simple_shuffling_sycl(const Eigen::SyclDevice& sycl_device)
     for (IndexType j = 0; j < sizeDim2; ++j) {
       for (IndexType k = 0; k < sizeDim3; ++k) {
         for (IndexType l = 0; l < sizeDim4; ++l) {
-          VERIFY_IS_EQUAL(tensor(i,j,k,l), no_shuffle(i,j,k,l));
+          VERIFY_IS_EQUAL(tensor(i, j, k, l), no_shuffle(i, j, k, l));
         }
       }
     }
@@ -78,12 +76,14 @@ static void test_simple_shuffling_sycl(const Eigen::SyclDevice& sycl_device)
   shuffles[1] = 3;
   shuffles[2] = 1;
   shuffles[3] = 0;
-  array<IndexType, 4> tensorrangeShuffle = {{sizeDim3, sizeDim4, sizeDim2, sizeDim1}};
-  Tensor<DataType, 4, DataLayout,IndexType> shuffle(tensorrangeShuffle);
-  DataType* gpu_data3  = static_cast<DataType*>(sycl_device.allocate(buffSize));
-  TensorMap<Tensor<DataType, 4,DataLayout,IndexType>> gpu3(gpu_data3, tensorrangeShuffle);
-
-  gpu3.device(sycl_device)=gpu1.shuffle(shuffles);
+  array<IndexType, 4> tensorrangeShuffle = {
+      {sizeDim3, sizeDim4, sizeDim2, sizeDim1}};
+  Tensor<DataType, 4, DataLayout, IndexType> shuffle(tensorrangeShuffle);
+  DataType* gpu_data3 = static_cast<DataType*>(sycl_device.allocate(buffSize));
+  TensorMap<Tensor<DataType, 4, DataLayout, IndexType>> gpu3(
+      gpu_data3, tensorrangeShuffle);
+
+  gpu3.device(sycl_device) = gpu1.shuffle(shuffles);
   sycl_device.memcpyDeviceToHost(shuffle.data(), gpu_data3, buffSize);
   sycl_device.synchronize();
 
@@ -96,24 +96,22 @@ static void test_simple_shuffling_sycl(const Eigen::SyclDevice& sycl_device)
     for (IndexType j = 0; j < sizeDim2; ++j) {
       for (IndexType k = 0; k < sizeDim3; ++k) {
         for (IndexType l = 0; l < sizeDim4; ++l) {
-          VERIFY_IS_EQUAL(tensor(i,j,k,l), shuffle(k,l,j,i));
+          VERIFY_IS_EQUAL(tensor(i, j, k, l), shuffle(k, l, j, i));
         }
       }
     }
   }
 }
 
-
-template<typename DataType, typename dev_Selector> void sycl_shuffling_test_per_device(dev_Selector s){
+template <typename DataType, typename dev_Selector>
+void sycl_shuffling_test_per_device(dev_Selector s) {
   QueueInterface queueInterface(s);
   auto sycl_device = Eigen::SyclDevice(&queueInterface);
   test_simple_shuffling_sycl<DataType, RowMajor, int64_t>(sycl_device);
   test_simple_shuffling_sycl<DataType, ColMajor, int64_t>(sycl_device);
-
 }
-EIGEN_DECLARE_TEST(cxx11_tensor_shuffling_sycl)
-{
-  for (const auto& device :Eigen::get_sycl_supported_devices()) {
+EIGEN_DECLARE_TEST(cxx11_tensor_shuffling_sycl) {
+  for (const auto& device : Eigen::get_sycl_supported_devices()) {
     CALL_SUBTEST(sycl_shuffling_test_per_device<float>(device));
   }
 }