Merge pull request #3363 from NervanaSystems/aprocter/logical-xor

Add logical xor op

doc/sphinx/source/ops/index.rst

@@ -77,6 +77,7 @@ Not currently a comprehensive list.
    * :doc:`tan`
    * :doc:`tanh`
    * :doc:`transpose`
+   * :doc:`xor`
 
 
 
@@ -149,6 +150,7 @@ Not currently a comprehensive list.
    tan.rst
    tanh.rst
    transpose.rst
+   xor.rst
 
 
 .. _more_about: 

doc/sphinx/source/ops/xor.rst

+.. xor.rst:
+
+###
+Xor
+###
+
+.. code-block:: cpp
+
+   Xor  // Elementwise logical-xor operation
+
+
+Description
+===========
+
+Produces tensor with boolean element type and shape as the two inputs,
+which must themselves have boolean element type, where the value at each
+coordinate of ``output`` is ``0`` (true) if ``arg0`` or ``arg1`` both
+zero or both nonzero, or ``1`` otherwise.
+
+
+Inputs
+------
+
++-----------------+------------------------------+--------------------------------+
+| Name            | Element Type                 | Shape                          |
++=================+==============================+================================+
+| ``arg0``        | ``ngraph::element::boolean`` | any                            |
++-----------------+------------------------------+--------------------------------+
+| ``arg1``        | ``ngraph::element::boolean`` | same as ``arg0``               |
++-----------------+------------------------------+--------------------------------+
+
+Outputs
+-------
+
++-----------------+------------------------------+--------------------------------+
+| Name            | Element Type                 | Shape                          |
++=================+==============================+================================+
+| ``output``      | ``ngraph::element::boolean`` | same as ``arg0``               |
++-----------------+------------------------------+--------------------------------+
+
+
+Mathematical Definition
+=======================
+
+.. math::
+
+   \mathtt{output}_{i_0, \ldots, i_{n-1}} = \mathtt{arg0}_{i_0, \ldots, i_{n-1}}\, \mathtt{XOR}\, \mathtt{arg1}_{i_0, \ldots, i_{n-1}}
+
+
+C++ Interface
+=============
+
+.. doxygenclass:: ngraph::op::Xor
+   :project: ngraph
+   :members:

src/ngraph/CMakeLists.txt

@@ -304,6 +304,8 @@ set (SRC
     op/tanh.hpp
     op/topk.cpp
     op/topk.hpp
+    op/xor.cpp
+    op/xor.hpp
     op/fused/clamp.cpp
     op/fused/clamp.hpp
     op/fused/conv_fused.cpp

src/ngraph/ngraph.hpp

@@ -175,6 +175,7 @@
 #include "ngraph/op/tanh.hpp"
 #include "ngraph/op/topk.hpp"
 #include "ngraph/op/util/attr_types.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/partial_shape.hpp"
 #include "ngraph/runtime/backend.hpp"
 #include "ngraph/runtime/tensor.hpp"

src/ngraph/op/op_tbl.hpp

@@ -160,3 +160,4 @@ NGRAPH_OP(Tanh, ngraph::op)
 NGRAPH_OP(Tile, ngraph::op)
 NGRAPH_OP(TopK, ngraph::op)
 NGRAPH_OP(Transpose, ngraph::op)
+NGRAPH_OP(Xor, ngraph::op)

src/ngraph/op/xor.cpp

+//*****************************************************************************
+// Copyright 2017-2019 Intel Corporation
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//*****************************************************************************
+
+#include "ngraph/op/xor.hpp"
+
+using namespace std;
+using namespace ngraph;
+
+const string op::Xor::type_name{"Xor"};
+
+op::Xor::Xor(const Output<Node>& arg0, const Output<Node>& arg1, const AutoBroadcastSpec& autob)
+    : BinaryElementwiseLogical(arg0, arg1, autob)
+{
+    constructor_validate_and_infer_types();
+}
+
+shared_ptr<Node> op::Xor::copy_with_new_args(const NodeVector& new_args) const
+{
+    check_new_args_count(this, new_args);
+    return make_shared<Xor>(new_args.at(0), new_args.at(1), this->get_autob());
+}

src/ngraph/op/xor.hpp

+//*****************************************************************************
+// Copyright 2017-2019 Intel Corporation
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//*****************************************************************************
+
+#pragma once
+
+#include <memory>
+
+#include "ngraph/op/util/binary_elementwise_logical.hpp"
+
+namespace ngraph
+{
+    namespace op
+    {
+        /// \brief Elementwise logical-xor operation.
+        ///
+        class Xor : public util::BinaryElementwiseLogical
+        {
+        public:
+            NGRAPH_API
+            static const std::string type_name;
+            const std::string& description() const override { return type_name; }
+            /// \brief Constructs a logical-xor operation.
+            ///
+            /// \param arg0 Node that produces the first input tensor.<br>
+            /// `[d0, ...]`
+            /// \param arg1 Node that produces the second input tensor.<br>
+            /// `[d0, ...]`
+            /// \param autob Auto broadcast specification
+            ///
+            /// Output `[d0, ...]`
+            ///
+            Xor(const Output<Node>& arg0,
+                const Output<Node>& arg1,
+                const AutoBroadcastSpec& autob = AutoBroadcastSpec());
+
+            virtual std::shared_ptr<Node>
+                copy_with_new_args(const NodeVector& new_args) const override;
+
+            virtual bool is_commutative() const override { return true; }
+        };
+    }
+}

src/ngraph/pass/constant_folding.cpp

@@ -65,6 +65,7 @@
 #include "ngraph/op/sqrt.hpp"
 #include "ngraph/op/subtract.hpp"
 #include "ngraph/op/sum.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/pattern/matcher.hpp"
 #include "ngraph/pattern/op/label.hpp"
 #include "ngraph/runtime/reference/abs.hpp"
@@ -107,6 +108,7 @@
 #include "ngraph/runtime/reference/sqrt.hpp"
 #include "ngraph/runtime/reference/subtract.hpp"
 #include "ngraph/runtime/reference/sum.hpp"
+#include "ngraph/runtime/reference/xor.hpp"
 #include "ngraph/slice_plan.hpp"
 #include "ngraph/util.hpp"
 
@@ -994,6 +996,17 @@ shared_ptr<op::Constant> fold_constant_binary(shared_ptr<op::Constant> a,
                                               shape_size(out_shape));
             return make_shared<op::Constant>(binary->get_element_type(), out_shape, out_vec);
         }
+        else if (std::dynamic_pointer_cast<op::Xor>(binary))
+        {
+            NGRAPH_CHECK(element::from<Tin>() == element::from<Tout>(),
+                         "Input/output types do not match");
+            vector<Tin> out_vec(shape_size(out_shape));
+            runtime::reference::logical_xor<Tin>(a->get_data_ptr<Tin>(),
+                                                 b->get_data_ptr<Tin>(),
+                                                 out_vec.data(),
+                                                 shape_size(out_shape));
+            return make_shared<op::Constant>(binary->get_element_type(), out_shape, out_vec);
+        }
         else
         {
             NGRAPH_CHECK(false,
@@ -1034,14 +1047,15 @@ shared_ptr<op::Constant> fold_constant_binary_helper(const element::Type& et_out
 }
 bool is_supported_binary_op(std::shared_ptr<Node> n)
 {
-    return (
-        std::dynamic_pointer_cast<op::Add>(n) || std::dynamic_pointer_cast<op::And>(n) ||
-        std::dynamic_pointer_cast<op::Divide>(n) || std::dynamic_pointer_cast<op::Equal>(n) ||
-        std::dynamic_pointer_cast<op::Greater>(n) || std::dynamic_pointer_cast<op::GreaterEq>(n) ||
-        std::dynamic_pointer_cast<op::Less>(n) || std::dynamic_pointer_cast<op::LessEq>(n) ||
-        std::dynamic_pointer_cast<op::Maximum>(n) || std::dynamic_pointer_cast<op::Minimum>(n) ||
-        std::dynamic_pointer_cast<op::Multiply>(n) || std::dynamic_pointer_cast<op::NotEqual>(n) ||
-        std::dynamic_pointer_cast<op::Or>(n) || std::dynamic_pointer_cast<op::Subtract>(n));
+    return (std::dynamic_pointer_cast<op::Add>(n) || std::dynamic_pointer_cast<op::And>(n) ||
+            std::dynamic_pointer_cast<op::Divide>(n) || std::dynamic_pointer_cast<op::Equal>(n) ||
+            std::dynamic_pointer_cast<op::Greater>(n) ||
+            std::dynamic_pointer_cast<op::GreaterEq>(n) || std::dynamic_pointer_cast<op::Less>(n) ||
+            std::dynamic_pointer_cast<op::LessEq>(n) || std::dynamic_pointer_cast<op::Maximum>(n) ||
+            std::dynamic_pointer_cast<op::Minimum>(n) ||
+            std::dynamic_pointer_cast<op::Multiply>(n) ||
+            std::dynamic_pointer_cast<op::NotEqual>(n) || std::dynamic_pointer_cast<op::Or>(n) ||
+            std::dynamic_pointer_cast<op::Subtract>(n) || std::dynamic_pointer_cast<op::Xor>(n));
 }
 
 void pass::ConstantFolding::construct_constant_binary()

src/ngraph/runtime/cpu/cpu_builder.cpp

@@ -65,6 +65,7 @@
 #include "ngraph/op/subtract.hpp"
 #include "ngraph/op/tan.hpp"
 #include "ngraph/op/tanh.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/runtime/cpu/cpu_builder_registry.hpp"
 #include "ngraph/runtime/cpu/cpu_kernels.hpp"
 #include "ngraph/runtime/cpu/cpu_op_annotations.hpp"
@@ -104,6 +105,7 @@
 #include "ngraph/runtime/cpu/kernel/subtract.hpp"
 #include "ngraph/runtime/cpu/kernel/tan.hpp"
 #include "ngraph/runtime/cpu/kernel/tanh.hpp"
+#include "ngraph/runtime/cpu/kernel/xor.hpp"
 #include "ngraph/runtime/cpu/op/convert_layout.hpp"
 #include "ngraph/runtime/cpu/op/halide_op.hpp"
 #include "ngraph/type/element_type.hpp"
@@ -243,6 +245,28 @@ namespace ngraph
                 functors.emplace_back(functor);
             }
 
+            template <>
+            void Builder::BUILDER_DECL(ngraph::op::Xor)
+            {
+                auto& functors = external_function->get_functors();
+
+                auto element_count = out[0].get_size();
+                auto arg0_buffer_index = external_function->get_buffer_index(args[0].get_name());
+                auto arg1_buffer_index = external_function->get_buffer_index(args[1].get_name());
+                auto out0_buffer_index = external_function->get_buffer_index(out[0].get_name());
+
+                auto functor =
+                    [&, element_count, arg0_buffer_index, arg1_buffer_index, out0_buffer_index](
+                        CPURuntimeContext* ctx, CPUExecutionContext* ectx) {
+                        runtime::cpu::kernel::logical_xor(ctx->buffer_data[arg0_buffer_index],
+                                                          ctx->buffer_data[arg1_buffer_index],
+                                                          ctx->buffer_data[out0_buffer_index],
+                                                          element_count,
+                                                          ectx->arena);
+                    };
+                functors.emplace_back(functor);
+            }
+
             template <>
             void Builder::BUILDER_DECL(ngraph::op::Maximum)
             {
@@ -545,6 +569,19 @@ namespace ngraph
                 return functor;
             }
 
+            template <>
+            NodeExecutorTy Builder::BUILDER_CF_DECL(ngraph::op::Xor)
+            {
+                auto element_count = shape_size(node->get_shape());
+
+                auto functor = [&, element_count](const std::vector<void*>& inputs,
+                                                  std::vector<void*>& outputs) {
+                    runtime::cpu::kernel::logical_xor(
+                        inputs[0], inputs[1], outputs[0], element_count, 0);
+                };
+                return functor;
+            }
+
             template <>
             NodeExecutorTy Builder::BUILDER_CF_DECL(ngraph::op::Sign)
             {
@@ -612,6 +649,7 @@ namespace ngraph
             REGISTER_OP_BUILDER(Minimum);
             REGISTER_OP_BUILDER(And);
             REGISTER_OP_BUILDER(Or);
+            REGISTER_OP_BUILDER(Xor);
 
             REGISTER_CF_BUILDER(Add);
             REGISTER_CF_BUILDER(Subtract);
@@ -633,6 +671,7 @@ namespace ngraph
             REGISTER_CF_BUILDER(LessEq);
             REGISTER_CF_BUILDER(And);
             REGISTER_CF_BUILDER(Or);
+            REGISTER_CF_BUILDER(Xor);
             REGISTER_CF_BUILDER(Sign);
             REGISTER_CF_BUILDER(Not);
         }

src/ngraph/runtime/cpu/cpu_emitter.cpp

@@ -113,6 +113,7 @@
 #include "ngraph/op/tan.hpp"
 #include "ngraph/op/tanh.hpp"
 #include "ngraph/op/topk.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/runtime/cpu/cpu_executor.hpp"
 #include "ngraph/runtime/cpu/cpu_kernel_emitters.hpp"
 #include "ngraph/runtime/cpu/cpu_op_annotations.hpp"
@@ -3822,6 +3823,15 @@ namespace ngraph
                        << "                      " << out[0].get_size() << ");\n";
             }
 
+            template <>
+            void CPU_Emitter::EMITTER_DECL(ngraph::op::Xor)
+            {
+                writer << "reference::logical_xor(" << args[0].get_name() << ",\n"
+                       << "                       " << args[1].get_name() << ",\n"
+                       << "                       " << out[0].get_name() << ",\n"
+                       << "                       " << out[0].get_size() << ");\n";
+            }
+
 #define TI(x) std::type_index(typeid(x))
 
             static std::string emit_infix_operator(const std::string& opname,

src/ngraph/runtime/cpu/cpu_external_function.cpp

@@ -134,6 +134,7 @@
 #include "ngraph/op/tan.hpp"
 #include "ngraph/op/tanh.hpp"
 #include "ngraph/op/topk.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/pass/algebraic_simplification.hpp"
 #include "ngraph/pass/batch_fusion.hpp"
 #include "ngraph/pass/common_function_collection.hpp"
@@ -431,6 +432,7 @@ static const runtime::cpu::OpMap dispatcher{
     {TI(ngraph::op::SigmoidBackprop), &runtime::cpu::CPU_Emitter::emit<op::SigmoidBackprop>},
     {TI(ngraph::op::And), &runtime::cpu::CPU_Emitter::emit<op::And>},
     {TI(ngraph::op::Or), &runtime::cpu::CPU_Emitter::emit<op::Or>},
+    {TI(ngraph::op::Xor), &runtime::cpu::CPU_Emitter::emit<op::Xor>},
     {TI(ngraph::op::CPULeakyRelu), &runtime::cpu::CPU_Emitter::emit<op::CPULeakyRelu>},
     {TI(ngraph::op::CompiledKernel), &runtime::cpu::CPU_Emitter::emit<op::CompiledKernel>},
     {TI(ngraph::op::LRN), &runtime::cpu::CPU_Emitter::emit<ngraph::op::LRN>},
@@ -567,6 +569,7 @@ void runtime::cpu::CPU_ExternalFunction::compile(ngraph::pass::PassConfig& pass_
 #include "ngraph/runtime/reference/slice.hpp"
 #include "ngraph/runtime/reference/sum.hpp"
 #include "ngraph/runtime/reference/topk.hpp"
+#include "ngraph/runtime/reference/xor.hpp"
 #include "ngraph/shape.hpp"
 #include "ngraph/state/rng_state.hpp"
 #include "ngraph/strides.hpp"

src/ngraph/runtime/cpu/kernel/xor.hpp

+//*****************************************************************************
+// Copyright 2017-2019 Intel Corporation
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//*****************************************************************************
+
+#pragma once
+
+#define EIGEN_USE_THREADS
+#include <unsupported/Eigen/CXX11/Tensor>
+
+#include "ngraph/runtime/cpu/cpu_executor.hpp"
+
+namespace ngraph
+{
+    namespace runtime
+    {
+        namespace cpu
+        {
+            namespace kernel
+            {
+                void logical_xor(void* input0, void* input1, void* output, size_t count, int arena)
+                {
+                    Eigen::array<Eigen::Index, 1> out_dims, in_dims;
+
+                    out_dims[0] = in_dims[0] = count;
+
+                    Eigen::TensorMap<Eigen::Tensor<char, 1, Eigen::RowMajor>> out(
+                        static_cast<char*>(output), out_dims);
+                    Eigen::TensorMap<Eigen::Tensor<char, 1, Eigen::RowMajor>> in0(
+                        static_cast<char*>(input0), in_dims);
+                    Eigen::TensorMap<Eigen::Tensor<char, 1, Eigen::RowMajor>> in1(
+                        static_cast<char*>(input1), in_dims);
+
+                    out.device(ngraph::runtime::cpu::executor::GetCPUExecutor().get_device(arena)) =
+                        (in0 != in1).template cast<char>();
+                }
+            }
+        }
+    }
+}

src/ngraph/runtime/generic_cpu/gcpu_executable.hpp

@@ -51,6 +51,7 @@
 #include "ngraph/op/max_pool.hpp"
 #include "ngraph/op/min.hpp"
 #include "ngraph/op/one_hot.hpp"
+#include "ngraph/op/or.hpp"
 #include "ngraph/op/pad.hpp"
 #include "ngraph/op/passthrough.hpp"
 #include "ngraph/op/product.hpp"
@@ -67,6 +68,7 @@
 #include "ngraph/op/softmax.hpp"
 #include "ngraph/op/sum.hpp"
 #include "ngraph/op/topk.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/runtime/aligned_buffer.hpp"
 #include "ngraph/runtime/backend.hpp"
 #include "ngraph/runtime/generic_cpu/kernel/broadcast.hpp"
@@ -154,6 +156,7 @@
 #include "ngraph/runtime/reference/tan.hpp"
 #include "ngraph/runtime/reference/tanh.hpp"
 #include "ngraph/runtime/reference/topk.hpp"
+#include "ngraph/runtime/reference/xor.hpp"
 #include "ngraph/runtime/tensor.hpp"
 #include "ngraph/state/rng_state.hpp"
 
@@ -1607,6 +1610,15 @@ private:
             }
             break;
         }
+        case OP_TYPEID::Xor:
+        {
+            size_t element_count = shape_size(node.get_output_shape(0));
+            reference::logical_xor(args[0]->get_data_ptr<const T>(),
+                                   args[1]->get_data_ptr<const T>(),
+                                   out[0]->get_data_ptr<T>(),
+                                   element_count);
+            break;
+        }
         case OP_TYPEID::DynBroadcast:
         case OP_TYPEID::Transpose:
         case OP_TYPEID::DynPad:

src/ngraph/runtime/gpu/gpu_emitter.cpp

@@ -129,6 +129,7 @@
 #include "ngraph/op/tan.hpp"
 #include "ngraph/op/tanh.hpp"
 #include "ngraph/op/topk.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/runtime/gpu/gpu_cuda_kernel_ops.hpp"
 #include "ngraph/runtime/gpu/gpu_emitter.hpp"
 #include "ngraph/runtime/gpu/gpu_kernel_emitters.hpp"
@@ -1463,6 +1464,11 @@ std::string runtime::gpu::GPU_Emitter::emit_TopK(EMIT_ARGS)
     return compiled_function->add_to_runtime(index, function_name, args, out);
 }
 
+std::string runtime::gpu::GPU_Emitter::emit_Xor(EMIT_ARGS)
+{
+    throw unsupported_op("Unsupported op '" + node->description() + "'");
+}
+
 std::string runtime::gpu::GPU_Emitter::emit_DynBroadcast(EMIT_ARGS)
 {
     throw unsupported_op("Unsupported op '" + node->description() + "'");

src/ngraph/runtime/gpu/unit_test.manifest

@@ -215,3 +215,4 @@ send_recv
 send_recv_ring
 gelu_f32
 gelu_f64
+logical_xor

src/ngraph/runtime/intelgpu/intelgpu_backend.cpp

@@ -2106,6 +2106,7 @@ shared_ptr<runtime::Executable>
         case OP_TYPEID::Tile:
         case OP_TYPEID::Transpose:
         case OP_TYPEID::Unsqueeze:
+        case OP_TYPEID::Xor:
         default:
         {
             throw unsupported_op("Unsupported op '" + op->description() +

src/ngraph/runtime/intelgpu/unit_test.manifest

@@ -112,6 +112,7 @@ send_recv
 send_recv_ring
 gelu_f32
 gelu_f64
+logical_xor
 
 # Not supported quant ops
 model_dequantize_linear_1d_zero_scale_int8

src/ngraph/runtime/interpreter/int_executable.hpp

@@ -154,6 +154,7 @@
 #include "ngraph/runtime/reference/tan.hpp"
 #include "ngraph/runtime/reference/tanh.hpp"
 #include "ngraph/runtime/reference/topk.hpp"
+#include "ngraph/runtime/reference/xor.hpp"
 #include "ngraph/runtime/tensor.hpp"
 #include "ngraph/state/rng_state.hpp"
 
@@ -1616,6 +1617,15 @@ private:
             }
             break;
         }
+        case OP_TYPEID::Xor:
+        {
+            size_t element_count = shape_size(node.get_output_shape(0));
+            reference::logical_xor(args[0]->get_data_ptr<const T>(),
+                                   args[1]->get_data_ptr<const T>(),
+                                   out[0]->get_data_ptr<T>(),
+                                   element_count);
+            break;
+        }
         case OP_TYPEID::DynBroadcast:
         case OP_TYPEID::Transpose:
         case OP_TYPEID::DynPad:

src/ngraph/runtime/plaidml/plaidml_ops_logical.cpp

@@ -17,6 +17,7 @@
 #include "ngraph/op/and.hpp"
 #include "ngraph/op/not.hpp"
 #include "ngraph/op/or.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/runtime/plaidml/plaidml_impl.hpp"
 
 namespace ngraph
@@ -28,6 +29,7 @@ namespace ngraph
             NGRAPH_PLAIDML_OP_CLASS(ImplAnd, OpImpl<op::And>);
             NGRAPH_PLAIDML_OP_CLASS(ImplNot, OpImpl<op::Not>);
             NGRAPH_PLAIDML_OP_CLASS(ImplOr, OpImpl<op::Or>);
+            NGRAPH_PLAIDML_OP_CLASS(ImplXor, OpImpl<op::Xor>);
         }
     }
 }
@@ -69,3 +71,16 @@ void ngraph::runtime::plaidml::ImplOr::Apply()
                    .add(builder::Elementwise{"C", "A ? A : B"})
                    .finalize());
 }
+
+// Xor performs a simple elementwise logical xor.
+void ngraph::runtime::plaidml::ImplXor::Apply()
+{
+    check_inputs(2);
+    check_outputs(1);
+    set_output(start_tile_function()
+                   .add(builder::Input{op_input(0), "A"})
+                   .add(builder::Input{op_input(1), "B"})
+                   .add(builder::Output{"C"})
+                   .add(builder::Elementwise{"C", "A ? (B ? 0 : A) : B"})
+                   .finalize());
+}

src/ngraph/runtime/plaidml/plaidml_pass_explicit_logicals.cpp

@@ -27,6 +27,7 @@
 #include "ngraph/op/not_equal.hpp"
 #include "ngraph/op/or.hpp"
 #include "ngraph/op/passthrough.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/pattern/matcher.hpp"
 #include "ngraph/pattern/op/any.hpp"
 #include "ngraph/pattern/op/any_of.hpp"
@@ -46,7 +47,8 @@ void ngraph::runtime::plaidml::pass::ExplicitLogicals::construct_logical_to_data
                 std::type_index{typeid(ngraph::op::LessEq)},
                 std::type_index{typeid(ngraph::op::Not)},
                 std::type_index{typeid(ngraph::op::NotEqual)},
-                std::type_index{typeid(ngraph::op::Or)}};
+                std::type_index{typeid(ngraph::op::Or)},
+                std::type_index{typeid(ngraph::op::Xor)}};
 
             const ngraph::Node* node_ptr = node.get();
 
@@ -62,7 +64,8 @@ void ngraph::runtime::plaidml::pass::ExplicitLogicals::construct_logical_to_data
                 std::type_index{typeid(ngraph::op::Equal)},
                 std::type_index{typeid(ngraph::op::Not)},
                 std::type_index{typeid(ngraph::op::NotEqual)},
-                std::type_index{typeid(ngraph::op::Or)}};
+                std::type_index{typeid(ngraph::op::Or)},
+                std::type_index{typeid(ngraph::op::Xor)}};
 
             const ngraph::Node* node_ptr = node.get();
 

src/ngraph/runtime/reference/xor.hpp

+//*****************************************************************************
+// Copyright 2017-2019 Intel Corporation
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//*****************************************************************************
+
+#pragma once
+
+#include <cstddef>
+
+namespace ngraph
+{
+    namespace runtime
+    {
+        namespace reference
+        {
+            template <typename T>
+            void logical_xor(const T* arg0, const T* arg1, T* out, size_t count)
+            {
+                for (size_t i = 0; i < count; i++)
+                {
+                    out[i] = static_cast<T>((arg0[i] || arg1[i]) && !(arg0[i] && arg1[i]));
+                }
+            }
+        }
+    }
+}

src/ngraph/serializer.cpp

@@ -146,6 +146,7 @@
 #include "ngraph/op/tan.hpp"
 #include "ngraph/op/tanh.hpp"
 #include "ngraph/op/topk.hpp"
+#include "ngraph/op/xor.hpp"
 #include "ngraph/provenance.hpp"
 #include "ngraph/serializer.hpp"
 #include "ngraph/util.hpp"
@@ -1920,6 +1921,11 @@ shared_ptr<Node> JSONDeserializer::deserialize_node(json node_js)
             node = make_shared<op::Unsqueeze>(args[0], args[1]);
             break;
         }
+        case OP_TYPEID::Xor:
+        {
+            node = make_shared<op::Xor>(args[0], args[1], read_auto_broadcast(node_js, "autob"));
+            break;
+        }
         case OP_TYPEID::UnknownOp:
         {
             stringstream ss;
@@ -2912,6 +2918,15 @@ json JSONSerializer::serialize_node(const Node& n)
     }
     case OP_TYPEID::Unsqueeze: { break;
     }
+    case OP_TYPEID::Xor:
+    {
+        auto tmp = dynamic_cast<const op::Xor*>(&n);
+        if (tmp->get_autob().m_type != op::AutoBroadcastType::NONE)
+        {
+            node["autob"] = write_auto_broadcast(tmp->get_autob());
+        }
+        break;
+    }
     case OP_TYPEID::UnknownOp: { break;
     }
     }

test/CMakeLists.txt

@@ -245,6 +245,7 @@ set(MULTI_TEST_SRC
     backend/generate_mask.in.cpp
     backend/logical_and.in.cpp
     backend/logical_or.in.cpp
+    backend/logical_xor.in.cpp
     backend/lrn.in.cpp
     backend/max.in.cpp
     backend/min.in.cpp

test/backend/logical_xor.in.cpp

+//*****************************************************************************
+// Copyright 2017-2019 Intel Corporation
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//*****************************************************************************
+
+#include "gtest/gtest.h"
+#include "ngraph/ngraph.hpp"
+#include "util/all_close.hpp"
+#include "util/all_close_f.hpp"
+#include "util/known_element_types.hpp"
+#include "util/ndarray.hpp"
+#include "util/test_control.hpp"
+#include "util/test_tools.hpp"
+
+using namespace std;
+using namespace ngraph;
+
+static string s_manifest = "${MANIFEST}";
+
+NGRAPH_TEST(${BACKEND_NAME}, logical_xor)
+{
+    Shape shape{2, 2, 2};
+    auto A = make_shared<op::Parameter>(element::boolean, shape);
+    auto B = make_shared<op::Parameter>(element::boolean, shape);
+    auto f = make_shared<Function>(make_shared<op::Xor>(A, B), ParameterVector{A, B});
+
+    auto backend = runtime::Backend::create("${BACKEND_NAME}");
+
+    // Create some tensors for input/output
+    auto a = backend->create_tensor(element::boolean, shape);
+    copy_data(a, vector<char>{1, 0, 1, 1, 1, 0, 1, 0});
+    auto b = backend->create_tensor(element::boolean, shape);
+    copy_data(b, vector<char>{0, 0, 1, 0, 0, 1, 1, 0});
+    auto result = backend->create_tensor(element::boolean, shape);
+
+    auto handle = backend->compile(f);
+    handle->call_with_validate({result}, {a, b});
+    EXPECT_EQ((vector<char>{1, 0, 0, 1, 1, 1, 0, 0}), read_vector<char>(result));
+}

test/constant_folding.cpp

@@ -822,6 +822,32 @@ TEST(constant_folding, const_or)
     ASSERT_EQ(values_expected, values_out);
 }
 
+TEST(constant_folding, const_xor)
+{
+    auto constant0 =
+        op::Constant::create(element::boolean, Shape{2, 3}, vector<int32_t>{0, 0, 1, 0, 1, 1});
+    auto constant1 =
+        op::Constant::create(element::boolean, Shape{2, 3}, vector<int32_t>{0, 1, 1, 1, 0, 1});
+    auto eq = make_shared<op::Xor>(constant0, constant1);
+    auto f = make_shared<Function>(eq, ParameterVector{});
+
+    pass::Manager pass_manager;
+    pass_manager.register_pass<pass::ConstantFolding>();
+    pass_manager.run_passes(f);
+
+    ASSERT_EQ(count_ops_of_type<op::Xor>(f), 0);
+    ASSERT_EQ(count_ops_of_type<op::Constant>(f), 1);
+
+    auto new_const =
+        std::dynamic_pointer_cast<op::Constant>(f->get_results().at(0)->get_argument(0));
+    ASSERT_TRUE(new_const);
+    auto values_out = new_const->get_vector<char>();
+
+    vector<char> values_expected{0, 1, 0, 1, 1, 0};
+
+    ASSERT_EQ(values_expected, values_out);
+}
+
 TEST(constant_folding, const_ceiling)
 {
     auto constant = op::Constant::create(

test/cpu_test.cpp

@@ -1216,14 +1216,16 @@ TEST(cpu_test, constant_unary_binary)
     auto less_eq = make_shared<op::LessEq>(g, h);
     auto logical_and = make_shared<op::And>(i, j);
     auto logical_or = make_shared<op::Or>(i, j);
+    auto logical_xor = make_shared<op::Xor>(i, j);
     auto ceil = make_shared<op::Ceiling>(k);
     auto floor = make_shared<op::Floor>(k);
     auto logical_not = make_shared<op::Not>(j);
 
     auto func = make_shared<Function>(
-        NodeVector{add,     sub,         mul,        divn,  min,       max,        absn,       neg,
-                   sqrt,    relu,        sign,       equal, not_equal, greater,    greater_eq, less,
-                   less_eq, logical_and, logical_or, ceil,  floor,     logical_not},
+        NodeVector{add,        sub,         mul,        divn,  min,        max,
+                   absn,       neg,         sqrt,       relu,  sign,       equal,
+                   not_equal,  greater,     greater_eq, less,  less_eq,    logical_and,
+                   logical_or, logical_xor, ceil,       floor, logical_not},
         ParameterVector{});
 
     auto func_error = make_shared<Function>(NodeVector{neg_sqrt}, ParameterVector{});
@@ -1252,6 +1254,7 @@ TEST(cpu_test, constant_unary_binary)
     ASSERT_EQ(count_ops_of_type<op::LessEq>(func), 0);
     ASSERT_EQ(count_ops_of_type<op::And>(func), 0);
     ASSERT_EQ(count_ops_of_type<op::Or>(func), 0);
+    ASSERT_EQ(count_ops_of_type<op::Xor>(func), 0);
     ASSERT_EQ(count_ops_of_type<op::Ceiling>(func), 0);
     ASSERT_EQ(count_ops_of_type<op::Floor>(func), 0);
     ASSERT_EQ(count_ops_of_type<op::Not>(func), 0);
@@ -1275,6 +1278,7 @@ TEST(cpu_test, constant_unary_binary)
     vector<char> less_eq_expected{1, 1, 1, 0};
     vector<char> and_expected{0, 0, 0, 1};
     vector<char> or_expected{0, 1, 1, 1};
+    vector<char> xor_expected{0, 1, 1, 0};
     vector<float> ceil_expected{0.0f, 0.0f, -1.0f, 3.0f};
     vector<float> floor_expected{-1.0f, 0.0f, -2.0f, 2.0f};
     vector<char> not_expected{1, 0, 1, 0};
@@ -1298,11 +1302,12 @@ TEST(cpu_test, constant_unary_binary)
     ASSERT_EQ(get_result_constant<char>(func, 16), less_eq_expected);
     ASSERT_EQ(get_result_constant<char>(func, 17), and_expected);
     ASSERT_EQ(get_result_constant<char>(func, 18), or_expected);
+    ASSERT_EQ(get_result_constant<char>(func, 19), xor_expected);
     ASSERT_TRUE(test::all_close_f(
-        get_result_constant<float>(func, 19), ceil_expected, MIN_FLOAT_TOLERANCE_BITS));
+        get_result_constant<float>(func, 20), ceil_expected, MIN_FLOAT_TOLERANCE_BITS));
     ASSERT_TRUE(test::all_close_f(
-        get_result_constant<float>(func, 20), floor_expected, MIN_FLOAT_TOLERANCE_BITS));
-    ASSERT_EQ(get_result_constant<char>(func, 21), not_expected);
+        get_result_constant<float>(func, 21), floor_expected, MIN_FLOAT_TOLERANCE_BITS));
+    ASSERT_EQ(get_result_constant<char>(func, 22), not_expected);
     ASSERT_ANY_THROW(pass_manager.run_passes(func_error));
 }
 

test/type_prop/binary_elementwise.cpp

@@ -209,6 +209,14 @@ TEST(type_prop, or_bad_arguments)
         });
 }
 
+TEST(type_prop, xor_bad_arguments)
+{
+    test_binary_logical(
+        "Xor", [](const shared_ptr<Node>& x, const shared_ptr<Node>& y) -> shared_ptr<Node> {
+            return make_shared<op::Xor>(x, y);
+        });
+}
+
 template <typename T>
 void test_binary_eltwise_numpy(const element::Type& et, const op::AutoBroadcastSpec& autob)
 {
@@ -242,6 +250,7 @@ TEST(type_prop, eltwise_auto_bcast)
     test_binary_eltwise_numpy<op::Or>(element::boolean, op::AutoBroadcastType::NUMPY);
     test_binary_eltwise_numpy<op::Power>(element::f32, op::AutoBroadcastType::NUMPY);
     test_binary_eltwise_numpy<op::Subtract>(element::f32, op::AutoBroadcastType::NUMPY);
+    test_binary_eltwise_numpy<op::Xor>(element::boolean, op::AutoBroadcastType::NUMPY);
 }
 
 TEST(type_prop, comparison_good)