IVGCVSW-3808 Add ElementwiseBinaryLayer

!android-nn-driver:9329

 * Added ElementwiseBinaryLayer that can represent all ElementwiseBinary
   operations including Add, Div, Sub, Maximum, Mul and Minimum.
 * Updated Delegate to use ElementwiseBinaryLayer instead of the Add,
   Div, Sub, Maximum, Mul and Minimum layers.
 * Updated Deserializer to use ElementwiseBinaryLayer instead of the Add,
   Div, Sub, Maximum, Mul and Minimum layers.
 * Updated OnnxParser to use ElementwiseBinaryLayer instead of the Add
   layer.
 * Updated TfLiteParser to use ElementwiseBinaryLayer instead of the Add,
   Div, Sub, Maximum, Mul and Minimum layers.
 * Updated CL and Neon tests to use ElementwiseBinaryLayer.
 * Updated CL and Neon Backend Specific Optimizations to accept
   ElementBinaryLayers as well as Add, Div, Mul, Sub, Maximum and Minimum
   layers.

Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: I7cbb96b60eb01f0e2b57b0541016d48a08b86c75

1 files changed, 54 insertions, 1 deletions

diff --git a/src/backends/neon/NeonWorkloadFactory.cpp b/src/backends/neon/NeonWorkloadFactory.cpp
index dccd4a3a36..08168eca2f 100644
--- a/src/backends/neon/NeonWorkloadFactory.cpp
+++ b/src/backends/neon/NeonWorkloadFactory.cpp
@@ -1,5 +1,5 @@
 //
-// Copyright © 2017 Arm Ltd and Contributors. All rights reserved.
+// Copyright © 2017-2023 Arm Ltd and Contributors. All rights reserved.
 // SPDX-License-Identifier: MIT
 //
 
@@ -280,6 +280,59 @@ std::unique_ptr<IWorkload> NeonWorkloadFactory::CreateWorkload(LayerType type,
             auto divisionQueueDescriptor = PolymorphicDowncast<const DivisionQueueDescriptor*>(&descriptor);
             return std::make_unique<NeonDivisionWorkload>(*divisionQueueDescriptor, info);
         }
+        case LayerType::ElementwiseBinary :
+        {
+            auto elementwiseBinaryQueueDescriptor
+                    = PolymorphicDowncast<const ElementwiseBinaryQueueDescriptor*>(&descriptor);
+
+            switch (elementwiseBinaryQueueDescriptor->m_Parameters.m_Operation)
+            {
+                case BinaryOperation::Add:
+                {
+                    AdditionQueueDescriptor additionQueueDescriptor;
+                    additionQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    additionQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonAdditionWorkload>(additionQueueDescriptor, info);
+                }
+                case BinaryOperation::Div:
+                {
+                    DivisionQueueDescriptor divisionQueueDescriptor;
+                    divisionQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    divisionQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonDivisionWorkload>(divisionQueueDescriptor, info);
+                }
+                case BinaryOperation::Maximum:
+                {
+                    MaximumQueueDescriptor maximumQueueDescriptor;
+                    maximumQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    maximumQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonMaximumWorkload>(maximumQueueDescriptor, info);
+                }
+                case BinaryOperation::Minimum:
+                {
+                    MinimumQueueDescriptor minimumQueueDescriptor;
+                    minimumQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    minimumQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonMinimumWorkload>(minimumQueueDescriptor, info);
+                }
+                case BinaryOperation::Mul:
+                {
+                    MultiplicationQueueDescriptor multiplicationQueueDescriptor;
+                    multiplicationQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    multiplicationQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonMultiplicationWorkload>(multiplicationQueueDescriptor, info);
+                }
+                case BinaryOperation::Sub:
+                {
+                    SubtractionQueueDescriptor subtractionQueueDescriptor;
+                    subtractionQueueDescriptor.m_Inputs = descriptor.m_Inputs;
+                    subtractionQueueDescriptor.m_Outputs = descriptor.m_Outputs;
+                    return std::make_unique<NeonSubtractionWorkload>(subtractionQueueDescriptor, info);
+                }
+                default:
+                    return nullptr;
+            }
+        }
         case LayerType::ElementwiseUnary :
         {
             auto elementwiseUnaryQueueDescriptor