From 9e53a35b66b1ec7ceee7c712380a13596175b83b Mon Sep 17 00:00:00 2001
From: arovir01 <Aron.Virginas-Tar@arm.com>
Date: Fri, 31 Aug 2018 15:26:35 +0100
Subject: IVGCVSW-1784: Rename float32 workloads for ACL

Change-Id: I98bdfe9cb12c663d1d5cfa456e2cc967d70ab22b
---
 .../NeonFullyConnectedFloatWorkload.cpp            | 96 ++++++++++++++++++++++
 1 file changed, 96 insertions(+)
 create mode 100644 src/armnn/backends/NeonWorkloads/NeonFullyConnectedFloatWorkload.cpp

(limited to 'src/armnn/backends/NeonWorkloads/NeonFullyConnectedFloatWorkload.cpp')

diff --git a/src/armnn/backends/NeonWorkloads/NeonFullyConnectedFloatWorkload.cpp b/src/armnn/backends/NeonWorkloads/NeonFullyConnectedFloatWorkload.cpp
new file mode 100644
index 0000000000..098d84bf0c
--- /dev/null
+++ b/src/armnn/backends/NeonWorkloads/NeonFullyConnectedFloatWorkload.cpp
@@ -0,0 +1,96 @@
+//
+// Copyright © 2017 Arm Ltd. All rights reserved.
+// See LICENSE file in the project root for full license information.
+//
+
+#include "NeonFullyConnectedFloatWorkload.hpp"
+
+#include "backends/ArmComputeTensorUtils.hpp"
+#include "backends/ArmComputeUtils.hpp"
+#include "backends/CpuTensorHandle.hpp"
+
+namespace armnn
+{
+using namespace armcomputetensorutils;
+
+arm_compute::Status NeonFullyConnectedWorkloadValidate(const TensorInfo& input,
+                                                       const TensorInfo& output,
+                                                       const TensorInfo& weights,
+                                                       const TensorInfo& biases,
+                                                       const FullyConnectedDescriptor& descriptor)
+{
+    const arm_compute::TensorInfo aclInput = BuildArmComputeTensorInfo(input);
+    const arm_compute::TensorInfo aclOutput = BuildArmComputeTensorInfo(output);
+    const arm_compute::TensorInfo aclWeights = BuildArmComputeTensorInfo(weights);
+
+    arm_compute::TensorInfo aclBiases;
+    arm_compute::TensorInfo *optionalAclBiases = nullptr;
+    if (descriptor.m_BiasEnabled)
+    {
+        aclBiases  = BuildArmComputeTensorInfo(biases);
+        optionalAclBiases = &aclBiases;
+    }
+
+    const arm_compute::FullyConnectedLayerInfo fullyConnectedLayerInfo =
+        ConvertFullyConnectedDescriptorToAclFullyConnectedLayerInfo(descriptor);
+
+
+    return arm_compute::NEFullyConnectedLayer::validate(&aclInput,
+                                                        &aclWeights,
+                                                        optionalAclBiases,
+                                                        &aclOutput,
+                                                        fullyConnectedLayerInfo);
+}
+
+NeonFullyConnectedFloatWorkload::NeonFullyConnectedFloatWorkload(const FullyConnectedQueueDescriptor& descriptor,
+    const WorkloadInfo& info, std::shared_ptr<arm_compute::MemoryManagerOnDemand>& memoryManager)
+    : FloatWorkload<FullyConnectedQueueDescriptor>(descriptor, info)
+    , m_FullyConnectedLayer(memoryManager)
+{
+    m_Data.ValidateInputsOutputs("NeonFullyConnectedFloatWorkload", 1, 1);
+
+    arm_compute::ITensor& input = boost::polymorphic_downcast<INeonTensorHandle*>(m_Data.m_Inputs[0])->GetTensor();
+    arm_compute::ITensor& output = boost::polymorphic_downcast<INeonTensorHandle*>(m_Data.m_Outputs[0])->GetTensor();
+
+    m_WeightsTensor = std::make_unique<arm_compute::Tensor>();
+    BuildArmComputeTensor(*m_WeightsTensor, m_Data.m_Weight->GetTensorInfo());
+
+    if (m_Data.m_Parameters.m_BiasEnabled)
+    {
+        m_BiasesTensor = std::make_unique<arm_compute::Tensor>();
+        BuildArmComputeTensor(*m_BiasesTensor, m_Data.m_Bias->GetTensorInfo());
+    }
+
+    // Construct
+    arm_compute::FullyConnectedLayerInfo fc_info;
+    fc_info.transpose_weights = m_Data.m_Parameters.m_TransposeWeightMatrix;
+    m_FullyConnectedLayer.configure(&input, m_WeightsTensor.get(), m_BiasesTensor.get(), &output, fc_info);
+
+    // Allocate
+    InitializeArmComputeTensorDataForFloatTypes(*m_WeightsTensor, m_Data.m_Weight);
+
+    if (m_BiasesTensor)
+    {
+        InitializeArmComputeTensorDataForFloatTypes(*m_BiasesTensor, m_Data.m_Bias);
+    }
+
+    // Force Compute Library to perform the necessary copying and reshaping, after which
+    // delete all the input tensors that will no longer be needed
+    m_FullyConnectedLayer.prepare();
+    FreeUnusedTensors();
+}
+
+void NeonFullyConnectedFloatWorkload::Execute() const
+{
+    ARMNN_SCOPED_PROFILING_EVENT_NEON("NeonFullyConnectedFloatWorkload_Execute");
+    m_FullyConnectedLayer.run();
+}
+
+void NeonFullyConnectedFloatWorkload::FreeUnusedTensors()
+{
+    FreeTensorIfUnused(m_WeightsTensor);
+    FreeTensorIfUnused(m_BiasesTensor);
+}
+
+} //namespace armnn
+
-- 
cgit v1.2.1