[Stride] Set up DenseTensorIterator And Support Stride Kernel For Elementwise_Add (#74637)

* add densetensor_iterator

* add HIP config

* set flag to true

* fix stride kernel bug

* add strided input test

* change flag name and add standard kernel defination

* refine

* fix codestyle

Author: Eddie-Wang1120

paddle/common/flags.cc

@@ -2180,3 +2180,15 @@ PHI_DEFINE_EXPORTED_bool(check_cuda_error,
 PHI_DEFINE_EXPORTED_bool(use_default_stream,
                          false,
                          "Whether use default stream.");
+
+/**
+ * Stride_Compute_Kernel related FLAG
+ * Name: FLAGS_use_stride_compute_kernel
+ * Since Version: 3.2
+ * Value Range: bool, default=false
+ * Example:
+ * Note: Whether use Stride_Compute_Kernel.
+ */
+PHI_DEFINE_EXPORTED_bool(use_stride_compute_kernel,
+                         false,
+                         "Whether use Stride_Compute_Kernel.");

paddle/phi/kernels/CMakeLists.txt

@@ -31,6 +31,7 @@ file(
   RELATIVE "${CMAKE_CURRENT_SOURCE_DIR}"
   "gpudnn/*.cu"
   "kps/*.cu"
+  "stride/*.cu"
   "legacy/kps/*.cu"
   "legacy/gpu/*.cu"
   "selected_rows/gpu/*.cu"