COMPMID-1768: NEON: Implement RealDiv

Change-Id: I0868669f7b733df141794fba1d79436e7581bd3a
Reviewed-on: https://review.mlplatform.org/426
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

3 files changed, 214 insertions, 2 deletions

diff --git a/tests/validation/NEON/ElementwiseDivision.cpp b/tests/validation/NEON/ElementwiseDivision.cpp
new file mode 100644
index 0000000000..86a2056ca8
--- /dev/null
+++ b/tests/validation/NEON/ElementwiseDivision.cpp
@@ -0,0 +1,155 @@
+/*
+ * Copyright (c) 2019 ARM Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "arm_compute/core/Types.h"
+#include "arm_compute/runtime/NEON/functions/NEElementwiseOperations.h"
+#include "arm_compute/runtime/Tensor.h"
+#include "arm_compute/runtime/TensorAllocator.h"
+#include "tests/NEON/Accessor.h"
+#include "tests/PaddingCalculator.h"
+#include "tests/datasets/ShapeDatasets.h"
+#include "tests/framework/Asserts.h"
+#include "tests/framework/Macros.h"
+#include "tests/framework/datasets/Datasets.h"
+#include "tests/validation/Validation.h"
+#include "tests/validation/fixtures/ElementwiseOperationsFixture.h"
+
+namespace arm_compute
+{
+namespace test
+{
+namespace validation
+{
+namespace
+{
+RelativeTolerance<float> tolerance_fp32(0.000001f);
+/** Input data sets **/
+#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
+RelativeTolerance<half> tolerance_fp16(static_cast<half>(0.01f));
+const auto              ElementwiseDivisionFP16Dataset = combine(combine(framework::dataset::make("DataType", DataType::F16), framework::dataset::make("DataType", DataType::F16)),
+                                                                 framework::dataset::make("DataType", DataType::F16));
+#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */
+const auto ElementwiseDivisionFP32Dataset = combine(combine(framework::dataset::make("DataType", DataType::F32), framework::dataset::make("DataType", DataType::F32)),
+                                                    framework::dataset::make("DataType", DataType::F32));
+} // namespace
+
+TEST_SUITE(NEON)
+TEST_SUITE(ElementwiseDivision)
+
+template <typename T>
+using NEElementwiseDivisionFixture = ElementwiseDivisionValidationFixture<Tensor, Accessor, NEElementwiseDivision, T>;
+
+// *INDENT-OFF*
+// clang-format off
+DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(
+               framework::dataset::make("Input1Info", { TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                        TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::F32),
+                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),     // Invalid data type combination
+                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),     // Mismatching shapes
+                                                      }),
+               framework::dataset::make("Input2Info",{ TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::S32),
+                                                       TensorInfo(TensorShape(48U, 11U, 2U), 1, DataType::F32),
+                                                     })),
+               framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),
+                                                       TensorInfo(TensorShape(48U, 11U, 2U), 1, DataType::F32),
+                                                     })),
+               framework::dataset::make("Expected", { true, true, true, false, false})),
+               input1_info, input2_info, output_info, expected)
+{
+    ARM_COMPUTE_EXPECT(bool(NEElementwiseDivision::validate(&input1_info.clone()->set_is_resizable(false), &input2_info.clone()->set_is_resizable(false), &output_info.clone()->set_is_resizable(false))) == expected, framework::LogLevel::ERRORS);
+}
+// clang-format on
+// *INDENT-ON*
+
+TEST_SUITE(Float)
+#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
+TEST_SUITE(F16)
+FIXTURE_DATA_TEST_CASE(RunSmall, NEElementwiseDivisionFixture<half>, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), ElementwiseDivisionFP16Dataset))
+{
+    // Validate output
+    validate(Accessor(_target), _reference, tolerance_fp16, 0.01);
+}
+TEST_SUITE_END() // F16
+#endif           /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */
+
+TEST_SUITE(F32)
+DATA_TEST_CASE(Configuration, framework::DatasetMode::ALL, framework::dataset::concat(datasets::SmallShapes(), datasets::LargeShapes()),
+               shape)
+{
+    // Create tensors
+    Tensor ref_src1 = create_tensor<Tensor>(shape, DataType::F32);
+    Tensor ref_src2 = create_tensor<Tensor>(shape, DataType::F32);
+    Tensor dst      = create_tensor<Tensor>(shape, DataType::F32);
+
+    // Create and Configure function
+    NEElementwiseDivision add;
+    add.configure(&ref_src1, &ref_src2, &dst);
+
+    // Validate valid region
+    const ValidRegion valid_region = shape_to_valid_region(shape);
+    validate(dst.info()->valid_region(), valid_region);
+}
+
+FIXTURE_DATA_TEST_CASE(RunSmall, NEElementwiseDivisionFixture<float>, framework::DatasetMode::PRECOMMIT, combine(datasets::SmallShapes(), ElementwiseDivisionFP32Dataset))
+{
+    // Validate output
+    validate(Accessor(_target), _reference, tolerance_fp32, 0.01);
+}
+
+FIXTURE_DATA_TEST_CASE(RunLarge, NEElementwiseDivisionFixture<float>, framework::DatasetMode::NIGHTLY, combine(datasets::LargeShapes(), ElementwiseDivisionFP32Dataset))
+{
+    // Validate output
+    validate(Accessor(_target), _reference, tolerance_fp32, 0.01);
+}
+
+template <typename T>
+using NEElementwiseDivisionBroadcastFixture = ElementwiseDivisionBroadcastValidationFixture<Tensor, Accessor, NEElementwiseDivision, T>;
+
+FIXTURE_DATA_TEST_CASE(RunSmallBroadcast, NEElementwiseDivisionBroadcastFixture<float>, framework::DatasetMode::PRECOMMIT, combine(datasets::SmallShapesBroadcast(),
+                       ElementwiseDivisionFP32Dataset))
+{
+    // Validate output
+    validate(Accessor(_target), _reference, tolerance_fp32, 0.01);
+}
+
+FIXTURE_DATA_TEST_CASE(RunLargeBroadcast, NEElementwiseDivisionBroadcastFixture<float>, framework::DatasetMode::NIGHTLY, combine(datasets::LargeShapesBroadcast(),
+                       ElementwiseDivisionFP32Dataset))
+{
+    // Validate output
+    validate(Accessor(_target), _reference, tolerance_fp32, 0.01);
+}
+TEST_SUITE_END() // F32
+TEST_SUITE_END() // Float
+
+TEST_SUITE_END() // ElementwiseDivision
+TEST_SUITE_END() // NEON
+} // namespace validation
+} // namespace test
+} // namespace arm_compute
diff --git a/tests/validation/fixtures/ElementwiseOperationsFixture.h b/tests/validation/fixtures/ElementwiseOperationsFixture.h
index 8b88f6a8fb..53030842e7 100644
--- a/tests/validation/fixtures/ElementwiseOperationsFixture.h
+++ b/tests/validation/fixtures/ElementwiseOperationsFixture.h
@@ -325,6 +325,63 @@ public:
                                                                                              qinfo0, qinfo1, qinfo_out);
     }
 };
+
+template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
+class ElementwiseDivisionBroadcastValidationFixture : public ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>
+{
+public:
+    template <typename...>
+    void setup(const TensorShape &shape0, const TensorShape &shape1, DataType data_type0, DataType data_type1, DataType output_data_type)
+    {
+        ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(ArithmeticOperation::DIV, shape0, shape1,
+                                                                                             data_type0, data_type1, output_data_type,
+                                                                                             QuantizationInfo(), QuantizationInfo(), QuantizationInfo());
+    }
+};
+
+template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
+class ElementwiseDivisionValidationFixture : public ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>
+{
+public:
+    template <typename...>
+    void setup(const TensorShape &shape, DataType data_type0, DataType data_type1, DataType output_data_type)
+    {
+        ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(ArithmeticOperation::DIV, shape, shape,
+                                                                                             data_type0, data_type1, output_data_type,
+                                                                                             QuantizationInfo(), QuantizationInfo(), QuantizationInfo());
+    }
+};
+
+template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
+class ElementwiseDivisionValidationQuantizedFixture : public ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>
+{
+public:
+    template <typename...>
+    void setup(const TensorShape &shape, DataType data_type0, DataType data_type1, DataType output_data_type,
+               QuantizationInfo qinfo0, QuantizationInfo qinfo1, QuantizationInfo qinfo_out)
+
+    {
+        ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(ArithmeticOperation::DIV, shape, shape,
+                                                                                             data_type0, data_type1, output_data_type,
+                                                                                             qinfo0, qinfo1, qinfo_out);
+    }
+};
+
+template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
+class ElementwiseDivisionQuantizedBroadcastValidationFixture : public ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>
+{
+public:
+    template <typename...>
+    void setup(const TensorShape &shape0, const TensorShape &shape1, DataType data_type0, DataType data_type1, DataType output_data_type,
+               QuantizationInfo qinfo0, QuantizationInfo qinfo1, QuantizationInfo qinfo_out)
+
+    {
+        ArithmeticOperationsGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(ArithmeticOperation::DIV, shape0, shape1,
+                                                                                             data_type0, data_type1, output_data_type,
+                                                                                             qinfo0, qinfo1, qinfo_out);
+    }
+};
+
 } // namespace validation
 } // namespace test
 } // namespace arm_compute
diff --git a/tests/validation/reference/ElementwiseOperations.cpp b/tests/validation/reference/ElementwiseOperations.cpp
index 1f0d42b26e..6d533edea5 100644
--- a/tests/validation/reference/ElementwiseOperations.cpp
+++ b/tests/validation/reference/ElementwiseOperations.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018 ARM Limited.
+ * Copyright (c) 2018-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -74,7 +74,7 @@ T arithm_op(ArithmeticOperation op, T src1, T src2, ConvertPolicy convert_policy
     }
 
     T result;
-    if(op == ArithmeticOperation::ADD || op == ArithmeticOperation::SUB)
+    if(op == ArithmeticOperation::ADD || op == ArithmeticOperation::SUB || op == ArithmeticOperation::DIV)
     {
         result = (convert_policy == ConvertPolicy::SATURATE) ? saturate_cast<T>(val) : static_cast<T>(val);
     }