Enable global pooling optimization on OpenCL

- Add loop unrolling on X and use POOL_X and POOL_Y defines for the for
  loop

Resolves COMPMID-4573

Change-Id: I33cb825cfb55912ccb0ab9d03bd33a3dab4c8b44
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5872
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 24 insertions, 0 deletions

diff --git a/tests/validation/Helpers.cpp b/tests/validation/Helpers.cpp
index 0f5d5c5101..237a5a517c 100644
--- a/tests/validation/Helpers.cpp
+++ b/tests/validation/Helpers.cpp
@@ -349,6 +349,30 @@ void add_padding_x(std::initializer_list<ITensor *> tensors, const DataLayout &d
     }
 }
 
+void add_padding_y(std::initializer_list<ITensor *> tensors, const DataLayout &data_layout)
+{
+    if(data_layout == DataLayout::NHWC)
+    {
+        constexpr unsigned int lower = 1U;
+        constexpr unsigned int upper = 4U;
+
+        std::uniform_int_distribution<unsigned int> distribution(lower, upper);
+        size_t                                      seed_offset = 0;
+
+        for(ITensor *tensor : tensors)
+        {
+            ARM_COMPUTE_ERROR_ON(!tensor->info()->is_resizable());
+
+            std::mt19937 gen(library->seed() + seed_offset++);
+
+            const unsigned int top    = distribution(gen);
+            const unsigned int bottom = distribution(gen);
+
+            tensor->info()->extend_padding(PaddingSize(top, 0U, bottom, 0U));
+        }
+    }
+}
+
 template void get_tile(const SimpleTensor<float> &in, SimpleTensor<float> &roi, const Coordinates &coord);
 template void get_tile(const SimpleTensor<half> &in, SimpleTensor<half> &roi, const Coordinates &coord);
 template void get_tile(const SimpleTensor<int> &in, SimpleTensor<int> &roi, const Coordinates &coord);