CPU Direct Execution: Implement And

src/ngraph/runtime/cpu/cpu_builder.cpp

@@ -80,6 +80,7 @@
 #include "ngraph/runtime/cpu/cpu_op_annotations.hpp"
 #include "ngraph/runtime/cpu/kernel/abs.hpp"
 #include "ngraph/runtime/cpu/kernel/add.hpp"
+#include "ngraph/runtime/cpu/kernel/and.hpp"
 #include "ngraph/runtime/cpu/kernel/broadcast.hpp"
 #include "ngraph/runtime/cpu/kernel/ceil.hpp"
 #include "ngraph/runtime/cpu/kernel/cwise_pow.hpp"
@@ -181,6 +182,40 @@ namespace ngraph
                 BUILD_BINARY_ELEMWISE_FUNCTOR(runtime::cpu::kernel::less_eq);
             }
 
+            template <>
+            void Builder::BUILDER_DECL(ngraph::op::And)
+            {
+                auto& functors = external_function->get_functors();
+                auto& tensor_data = external_function->get_tensor_data();
+
+                auto element_count = out[0].get_size();
+                auto& arg0_tensor = tensor_data[args[0].get_name()];
+                auto& arg1_tensor = tensor_data[args[1].get_name()];
+                auto& out0_tensor = tensor_data[out[0].get_name()];
+
+                auto functor = [&, element_count](CPURuntimeContext* ctx) {
+                                   runtime::cpu::kernel::logical_and(arg0_tensor, arg1_tensor, out0_tensor, element_count);
+                               };
+                functors.emplace_back(functor);
+            }
+
+            template <>
+            void Builder::BUILDER_DECL(ngraph::op::Or)
+            {
+                auto& functors = external_function->get_functors();
+                auto& tensor_data = external_function->get_tensor_data();
+
+                auto element_count = out[0].get_size();
+                auto& arg0_tensor = tensor_data[args[0].get_name()];
+                auto& arg1_tensor = tensor_data[args[1].get_name()];
+                auto& out0_tensor = tensor_data[out[0].get_name()];
+
+                auto functor = [&, element_count](CPURuntimeContext* ctx) {
+                                   runtime::cpu::kernel::logical_or(arg0_tensor, arg1_tensor, out0_tensor, element_count);
+                               };
+                functors.emplace_back(functor);
+            }
+
             template <>
             void Builder::BUILDER_DECL(ngraph::op::Maximum)
             {
@@ -322,6 +357,8 @@ namespace ngraph
             REGISTER_OP_BUILDER(LessEq);
             REGISTER_OP_BUILDER(Maximum);
             REGISTER_OP_BUILDER(Minimum);
+            REGISTER_OP_BUILDER(And);
+            REGISTER_OP_BUILDER(Or);
         }
     }
 }

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

+/*******************************************************************************
+* Copyright 2018 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/kernel/eigen_thread_pool.hpp"
+
+namespace ngraph
+{
+    namespace runtime
+    {
+        namespace cpu
+        {
+            namespace kernel
+            {
+                void logical_and(void* input0, void* input1, void* output, size_t count)
+                {
+                    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(eigen::global_thread_pool_device) =
+                        (in0 && in1).template cast<char>();
+                }
+            }
+        }
+    }
+}