[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).

This changes moves the remaining modules in "core/graph/..." that depend on modules in "core/common_runtime/..." to "core/common_runtime/", and creates separate fine-grained targets for each of them:

* "gradients.{h.cc}"
* "graph_def_builder_util.{h,cc}"
* "mkl_{layout,tfconversion}_pass.{h,cc}"
* "quantize_training.{h,cc}"

PiperOrigin-RevId: 308841156
Change-Id: I7774e2498f465290e7addd331625bd91d90acbac

tensorflow/cc/BUILD

@@ -182,6 +182,7 @@ cc_library_with_android_deps(
     deps = [
         "//tensorflow/core:core_cpu",
         "//tensorflow/core:framework",
+        "//tensorflow/core:graph",
         "//tensorflow/core:lib",
         "//tensorflow/core:lib_internal",
         "//tensorflow/core:ops",

tensorflow/compiler/jit/cluster_scoping_pass_test.cc

@@ -18,12 +18,12 @@ limitations under the License.
 #include "absl/container/flat_hash_map.h"
 #include "tensorflow/compiler/jit/defs.h"
 #include "tensorflow/compiler/jit/test_util.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 #include "tensorflow/core/public/session_options.h"

tensorflow/compiler/jit/compilability_check_util_test.cc

@@ -21,12 +21,12 @@ limitations under the License.
 #include "tensorflow/cc/ops/standard_ops.h"
 #include "tensorflow/compiler/tf2xla/xla_op_kernel.h"
 #include "tensorflow/compiler/tf2xla/xla_op_registry.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/attr_value.pb.h"
 #include "tensorflow/core/framework/function.h"
 #include "tensorflow/core/framework/graph_to_functiondef.h"
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/compiler/jit/deadness_analysis_test.cc

@@ -30,7 +30,6 @@ limitations under the License.
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/compiler/jit/mark_for_compilation_pass_test.cc

@@ -13,8 +13,6 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
-#include "tensorflow/compiler/jit/mark_for_compilation_pass_test_helper.h"
-
 #include "absl/container/flat_hash_map.h"
 #include "absl/memory/memory.h"
 #include "absl/strings/match.h"
@@ -28,15 +26,16 @@ limitations under the License.
 #include "tensorflow/cc/ops/sendrecv_ops.h"
 #include "tensorflow/cc/ops/standard_ops.h"
 #include "tensorflow/compiler/jit/defs.h"
+#include "tensorflow/compiler/jit/mark_for_compilation_pass_test_helper.h"
 #include "tensorflow/compiler/jit/node_matchers.h"
 #include "tensorflow/compiler/tf2xla/xla_op_kernel.h"
 #include "tensorflow/compiler/tf2xla/xla_op_registry.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/compiler/jit/partially_decluster_pass_test.cc

@@ -28,6 +28,7 @@ limitations under the License.
 #include "tensorflow/compiler/tf2xla/cc/ops/xla_ops.h"
 #include "tensorflow/compiler/tf2xla/xla_op_kernel.h"
 #include "tensorflow/compiler/tf2xla/xla_op_registry.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/function.h"
 #include "tensorflow/core/framework/function.pb.h"
 #include "tensorflow/core/framework/node_def_util.h"
@@ -35,7 +36,6 @@ limitations under the License.
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/compiler/jit/resource_operation_safety_analysis_test.cc

@@ -26,12 +26,11 @@ limitations under the License.
 #include "tensorflow/compiler/jit/defs.h"
 #include "tensorflow/compiler/tf2xla/xla_op_kernel.h"
 #include "tensorflow/compiler/tf2xla/xla_op_registry.h"
+#include "tensorflow/core/common_runtime/graph_constructor.h"
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/graph/algorithm.h"
-#include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/BUILD

@@ -2796,7 +2796,6 @@ tf_cc_test_mkl(
     name = "mkl_related_tests",
     size = "small",
     srcs = [
-        "//tensorflow/core/graph:mkl_related_tests",
         "//tensorflow/core/util:mkl_util_test_srcs",
     ],
     linkstatic = 1,

tensorflow/core/common_runtime/BUILD

@@ -58,7 +58,9 @@ tf_cuda_library(
         "device_factory.h",
         "function.h",
         "function_optimization_registry.h",
+        "gradients.h",
         "graph_constructor.h",
+        "graph_def_builder_util.h",
         "optimization_registry.h",
         "shape_refiner.h",
         "//tensorflow/core/graph:core_cpu_headers",
@@ -157,6 +159,7 @@ filegroup(
         "function_body.h",
         "function_utils.h",
         "graph_constructor.h",
+        "graph_def_builder_util.h",
         "graph_optimizer.h",
         "graph_runner.h",
         "inline_function_utils.h",
@@ -173,7 +176,6 @@ filegroup(
 tf_cuda_library(
     name = "core_cpu_base_no_ops",
     srcs = [
-        "//tensorflow/core/graph:core_cpu_base_no_ops_srcs",
         "//tensorflow/core/public:session_options.h",
         "//tensorflow/core/public:version.h",
     ],
@@ -194,6 +196,7 @@ tf_cuda_library(
         "//third_party/eigen3",
     ] + if_static([
         ":graph_constructor",
+        ":graph_def_builder_util",
         "@com_google_absl//absl/algorithm:container",
         "@com_google_absl//absl/strings",
     ]),
@@ -225,6 +228,8 @@ filegroup(
         "executor.h",
         "executor_factory.h",
         "function_optimization_registry.h",
+        "graph_optimizer.h",
+        "gradients.h",
         "input_colocation_exemption_registry.h",
         "isolate_placer_inspection_required_ops_pass.h",
         "local_device.h",
@@ -236,12 +241,14 @@ filegroup(
         "lower_while_op.h",
         "memory_types.h",
         "mkl_cpu_allocator.h",
+        "mkl_layout_pass.h",
         "optimization_registry.h",
         "partitioning_utils.h",
         "placer.h",
         "process_util.h",
         "inspecting_placer.h",
         "profile_handler.h",
+        "quantize_training.h",
         "renamed_device.h",
         "rendezvous_mgr.h",
         "rendezvous_util.h",
@@ -256,7 +263,6 @@ filegroup(
         "threadpool_device.h",
         "process_state.h",
         "pool_allocator.h",
-        "//tensorflow/core/graph:core_cpu_lib_headers",
     ] + if_mkl(["//tensorflow/core/graph:mkl_graph_util_header"]),
 )
 
@@ -606,6 +612,23 @@ cc_library(
     ],
 )
 
+cc_library(
+    name = "gradients",
+    srcs = ["gradients.cc"],
+    hdrs = ["gradients.h"],
+    deps = [
+        ":device",
+        ":executor",
+        ":graph_constructor",
+        ":graph_optimizer",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:graph",
+        "//tensorflow/core:lib",
+        "//tensorflow/core:lib_internal",
+        "//tensorflow/core:protos_all_cc",
+    ],
+)
+
 # This library also includes "eval_const_tensor", "graph_runner", and
 # "shape_refiner", because there are circular dependencies between these
 # modules.
@@ -646,6 +669,18 @@ cc_library(
     ],
 )
 
+cc_library(
+    name = "graph_def_builder_util",
+    srcs = ["graph_def_builder_util.cc"],
+    hdrs = ["graph_def_builder_util.h"],
+    copts = tf_copts(),
+    deps = [
+        ":graph_constructor",
+        "//tensorflow/core:graph",
+        "//tensorflow/core:lib",
+    ],
+)
+
 cc_library(
     name = "graph_optimizer",
     srcs = ["graph_optimizer.cc"],
@@ -793,6 +828,36 @@ cc_library(
     ] + mkl_deps(),
 )
 
+cc_library(
+    name = "mkl_layout_pass",
+    srcs = ["mkl_layout_pass.cc"],
+    hdrs = ["mkl_layout_pass.h"],
+    copts = tf_copts(),
+    deps = [
+        ":optimization_registry",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:framework_internal",
+        "//tensorflow/core:graph",
+        "//tensorflow/core:lib",
+    ] + mkl_deps(),
+    alwayslink = 1,
+)
+
+cc_library(
+    name = "mkl_tfconversion_pass",
+    srcs = ["mkl_tfconversion_pass.cc"],
+    hdrs = ["mkl_tfconversion_pass.h"],
+    copts = tf_copts(),
+    deps = [
+        ":optimization_registry",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:framework_internal",
+        "//tensorflow/core:graph",
+        "//tensorflow/core:lib",
+    ] + mkl_deps(),
+    alwayslink = 1,
+)
+
 cc_library(
     name = "optimization_registry",
     srcs = ["optimization_registry.cc"],
@@ -923,6 +988,32 @@ cc_library(
     ],
 )
 
+filegroup(
+    name = "quantize_training_hdrs",
+    srcs = [
+        "quantize_training.h",
+    ],
+    visibility = [
+        "//tensorflow/python:__pkg__",
+    ],
+)
+
+cc_library(
+    name = "quantize_training",
+    srcs = ["quantize_training.cc"],
+    hdrs = [":quantize_training_hdrs"],
+    copts = tf_copts(),
+    deps = [
+        ":executor",
+        ":graph_constructor",
+        ":memory_types",
+        ":session_options",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:graph",
+        "//tensorflow/core:lib",
+    ],
+)
+
 cc_library(
     name = "renamed_device",
     srcs = ["renamed_device.cc"],
@@ -1215,7 +1306,6 @@ tf_cuda_library(
         "placer_inspection_required_ops_utils.cc",
         "placer_inspection_required_ops_utils.h",
         "process_function_library_runtime.cc",
-        "//tensorflow/core/graph:core_cpu_impl_srcs",
     ],
     hdrs = [":core_cpu_lib_headers"],
     copts = tf_copts(),
@@ -1226,6 +1316,7 @@ tf_cuda_library(
         ":executor_factory",
         ":function_body",
         ":function_optimization_registry",
+        ":gradients",
         ":graph_constructor",
         ":graph_optimizer",
         ":graph_view",
@@ -1285,12 +1376,15 @@ tf_cuda_library(
         ":device_resolver_local",
         ":device_set",
         ":entry",
+        ":graph_def_builder_util",
         ":graph_view",
         ":hierarchical_tree_broadcaster",
         ":input_colocation_exemption_registry",
         ":local_device",
         ":memory_types",
         ":mkl_cpu_allocator",
+        ":mkl_layout_pass",
+        ":mkl_tfconversion_pass",
         ":optimization_registry",
         ":parallel_concat_optimizer",
         ":partitioning_utils",
@@ -1299,6 +1393,7 @@ tf_cuda_library(
         ":process_state",
         ":process_util",
         ":profile_handler",
+        ":quantize_training",
         ":renamed_device",
         ":rendezvous_mgr",
         ":rendezvous_util",
@@ -2268,3 +2363,80 @@ tf_cc_test(
         "//tensorflow/core:testlib",
     ],
 )
+
+tf_cc_test_mkl(
+    name = "mkl_related_tests",
+    size = "small",
+    srcs = [
+        "mkl_layout_pass_test.cc",
+        "mkl_tfconversion_pass_test.cc",
+    ],
+    linkstatic = 1,
+    deps = [
+        ":core",
+        ":core_cpu",
+        ":core_cpu_internal",
+        ":direct_session_internal",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:framework_internal",
+        "//tensorflow/core:lib",
+        "//tensorflow/core:lib_internal",
+        "//tensorflow/core:ops",
+        "//tensorflow/core:protos_all_cc",
+        "//tensorflow/core:test",
+        "//tensorflow/core:test_main",
+        "//tensorflow/core:testlib",
+        "//tensorflow/cc:cc_ops",
+        "//tensorflow/cc:scope",
+        "//tensorflow/cc:sendrecv_ops",
+        "//tensorflow/core/kernels:ops_util",
+        "//third_party/eigen3",
+    ] + if_mkl([
+        "//tensorflow/core/kernels:mkl_aggregate_ops",
+        "//tensorflow/core/kernels:mkl_batch_matmul_op",
+        "//tensorflow/core/kernels:mkl_concat_op",
+        "//tensorflow/core/kernels:mkl_conv_op",
+        "//tensorflow/core/kernels:mkl_cwise_ops_common",
+        "//tensorflow/core/kernels:mkl_dequantize_op",
+        "//tensorflow/core/kernels:mkl_fused_batch_norm_op",
+        "//tensorflow/core/kernels:mkl_identity_op",
+        "//tensorflow/core/kernels:mkl_input_conversion_op",
+        "//tensorflow/core/kernels:mkl_lrn_op",
+        "//tensorflow/core/kernels:mkl_matmul_op",
+        "//tensorflow/core/kernels:mkl_pooling_ops",
+        "//tensorflow/core/kernels:mkl_qmatmul_op",
+        "//tensorflow/core/kernels:mkl_quantize_op",
+        "//tensorflow/core/kernels:mkl_relu_op",
+        "//tensorflow/core/kernels:mkl_reshape_op",
+        "//tensorflow/core/kernels:mkl_slice_op",
+        "//tensorflow/core/kernels:mkl_softmax_op",
+        "//tensorflow/core/kernels:mkl_tfconv_op",
+        "//tensorflow/core/kernels:mkl_transpose_op",
+        "//tensorflow/core/kernels:mkl_tmp_bf16_ops",
+    ]),
+)
+
+# TODO(bmzhao): Refactor this target to use granular dependencies
+# after stage 4 of the TF build refactor is complete:
+# https://github.com/tensorflow/community/pull/179
+tf_cc_test(
+    name = "quantize_training_test",
+    srcs = ["quantize_training_test.cc"],
+    deps = [
+        ":core_cpu",
+        ":core_cpu_internal",
+        ":direct_session_internal",
+        "//tensorflow/core",
+        "//tensorflow/core:all_kernels",
+        "//tensorflow/core:framework",
+        "//tensorflow/core:framework_internal",
+        "//tensorflow/core:lib",
+        "//tensorflow/core:lib_internal",
+        "//tensorflow/core:ops",
+        "//tensorflow/core:protos_all_cc",
+        "//tensorflow/core:test",
+        "//tensorflow/core:test_main",
+        "//tensorflow/core:testlib",
+        "//tensorflow/core/util:protos_test_cc",
+    ],
+)

tensorflow/core/common_runtime/eager/mkl_eager_op_rewrite.cc

@@ -17,8 +17,8 @@ limitations under the License.
 #include <unordered_map>
 
 #include "tensorflow/core/common_runtime/eager/eager_op_rewrite_registry.h"
+#include "tensorflow/core/common_runyime/mkl_layout_pass.h"
 #include "tensorflow/core/graph/mkl_graph_util.h"
-#include "tensorflow/core/graph/mkl_layout_pass.h"
 #include "tensorflow/core/lib/core/status.h"
 #include "tensorflow/core/util/mkl_util.h"
 #include "tensorflow/core/util/util.h"

tensorflow/core/common_runtime/function.cc

@@ -24,6 +24,7 @@ limitations under the License.
 #include "tensorflow/core/common_runtime/device.h"
 #include "tensorflow/core/common_runtime/executor.h"
 #include "tensorflow/core/common_runtime/executor_factory.h"
+#include "tensorflow/core/common_runtime/gradients.h"
 #include "tensorflow/core/common_runtime/graph_constructor.h"
 #include "tensorflow/core/common_runtime/graph_optimizer.h"
 #include "tensorflow/core/common_runtime/inline_function_utils.h"
@@ -38,7 +39,6 @@ limitations under the License.
 #include "tensorflow/core/framework/versions.pb.h"
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/control_flow.h"
-#include "tensorflow/core/graph/gradients.h"
 #include "tensorflow/core/graph/node_builder.h"
 #include "tensorflow/core/graph/optimizer_cse.h"
 #include "tensorflow/core/lib/core/threadpool.h"

tensorflow/core/common_runtime/gradients.cc

@@ -13,11 +13,14 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/common_runtime/gradients.h"
+
 #include <deque>
 #include <vector>
 
 #include "tensorflow/core/common_runtime/device.h"
 #include "tensorflow/core/common_runtime/executor.h"
+#include "tensorflow/core/common_runtime/graph_constructor.h"
 #include "tensorflow/core/common_runtime/graph_optimizer.h"
 #include "tensorflow/core/framework/function.h"
 #include "tensorflow/core/framework/node_def.pb.h"
@@ -25,8 +28,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/op_kernel.h"
 #include "tensorflow/core/graph/algorithm.h"
-#include "tensorflow/core/graph/gradients.h"
-#include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/optimizer_cse.h"
 #include "tensorflow/core/lib/gtl/map_util.h"
 #include "tensorflow/core/platform/macros.h"

tensorflow/core/common_runtime/gradients.h

@@ -13,8 +13,8 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
-#ifndef TENSORFLOW_CORE_GRAPH_GRADIENTS_H_
-#define TENSORFLOW_CORE_GRAPH_GRADIENTS_H_
+#ifndef TENSORFLOW_CORE_COMMON_RUNTIME_GRADIENTS_H_
+#define TENSORFLOW_CORE_COMMON_RUNTIME_GRADIENTS_H_
 
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/lib/core/status.h"
@@ -55,4 +55,4 @@ Status AddSymbolicGradients(gtl::ArraySlice<NodeOut> y_node_outputs,
 
 }  // namespace tensorflow
 
-#endif  // TENSORFLOW_CORE_GRAPH_GRADIENTS_H_
+#endif  // TENSORFLOW_CORE_COMMON_RUNTIME_GRADIENTS_H_

tensorflow/core/common_runtime/graph_def_builder_util.cc

@@ -12,9 +12,9 @@ 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 "tensorflow/core/graph/graph_def_builder_util.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 
-#include "tensorflow/core/graph/graph_constructor.h"
+#include "tensorflow/core/common_runtime/graph_constructor.h"
 
 namespace tensorflow {
 

tensorflow/core/common_runtime/graph_def_builder_util.h

@@ -12,8 +12,8 @@ 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.
 ==============================================================================*/
-#ifndef TENSORFLOW_CORE_GRAPH_GRAPH_DEF_BUILDER_UTIL_H_
-#define TENSORFLOW_CORE_GRAPH_GRAPH_DEF_BUILDER_UTIL_H_
+#ifndef TENSORFLOW_CORE_COMMON_RUNTIME_GRAPH_DEF_BUILDER_UTIL_H_
+#define TENSORFLOW_CORE_COMMON_RUNTIME_GRAPH_DEF_BUILDER_UTIL_H_
 
 #include "tensorflow/core/graph/graph_def_builder.h"
 #include "tensorflow/core/lib/core/status.h"
@@ -32,4 +32,4 @@ Status GraphDefBuilderToGraph(const GraphDefBuilder& builder, Graph* graph);
 
 }  // namespace tensorflow
 
-#endif  // TENSORFLOW_CORE_GRAPH_GRAPH_DEF_BUILDER_UTIL_H_
+#endif  // TENSORFLOW_CORE_COMMON_RUNTIME_GRAPH_DEF_BUILDER_UTIL_H_

tensorflow/core/common_runtime/isolate_placer_inspection_required_ops_pass_test.cc

@@ -27,7 +27,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 #include "tensorflow/core/util/equal_graph_def.h"

tensorflow/core/common_runtime/lower_case_op_test.cc

@@ -28,7 +28,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/common_runtime/lower_function_call_op_test.cc

@@ -28,7 +28,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/common_runtime/lower_functional_ops_test.cc

@@ -28,7 +28,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/common_runtime/lower_if_op_test.cc

@@ -28,7 +28,6 @@ limitations under the License.
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/common_runtime/lower_while_op_test.cc

@@ -26,7 +26,6 @@ limitations under the License.
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/common_runtime/mkl_layout_pass.cc

@@ -17,7 +17,7 @@ limitations under the License.
 // all over the place, we should log an error and execute the original graph.
 #ifdef INTEL_MKL
 
-#include "tensorflow/core/graph/mkl_layout_pass.h"
+#include "tensorflow/core/common_runtime/mkl_layout_pass.h"
 
 #include <algorithm>
 #include <functional>

tensorflow/core/common_runtime/mkl_layout_pass.h

@@ -15,8 +15,8 @@ limitations under the License.
 
 // A graph pass that rewrites graph for propagating MKL layout as a tensor
 
-#ifndef TENSORFLOW_CORE_GRAPH_MKL_LAYOUT_PASS_H_
-#define TENSORFLOW_CORE_GRAPH_MKL_LAYOUT_PASS_H_
+#ifndef TENSORFLOW_CORE_COMMON_RUNTIME_MKL_LAYOUT_PASS_H_
+#define TENSORFLOW_CORE_COMMON_RUNTIME_MKL_LAYOUT_PASS_H_
 
 #ifdef INTEL_MKL
 
@@ -33,4 +33,4 @@ extern bool RunMklLayoutRewritePass(std::unique_ptr<Graph>* g);
 
 #endif
 
-#endif  // TENSORFLOW_CORE_GRAPH_MKL_LAYOUT_PASS_H_
+#endif  // TENSORFLOW_CORE_COMMON_RUNTIME_MKL_LAYOUT_PASS_H_

tensorflow/core/common_runtime/mkl_layout_pass_test.cc

@@ -15,15 +15,15 @@ limitations under the License.
 
 #if defined(INTEL_MKL) && defined(ENABLE_MKL)
 
-#include "tensorflow/core/graph/mkl_layout_pass.h"
+#include "tensorflow/core/common_runtime/mkl_layout_pass.h"
 
 #include <algorithm>
 #include <vector>
 
+#include "tensorflow/core/common_runtime/graph_constructor.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor.h"
 #include "tensorflow/core/graph/graph.h"
-#include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/mkl_graph_util.h"
 #include "tensorflow/core/graph/testlib.h"
 #include "tensorflow/core/kernels/ops_util.h"

tensorflow/core/common_runtime/mkl_tfconversion_pass.cc

@@ -15,6 +15,8 @@ limitations under the License.
 
 #ifdef INTEL_MKL
 
+#include "tensorflow/core/common_runtime/mkl_tfconversion_pass.h"
+
 #include <memory>
 #include <queue>
 #include <set>
@@ -26,6 +28,7 @@ limitations under the License.
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/graph/algorithm.h"
 #include "tensorflow/core/graph/graph.h"
+#include "tensorflow/core/graph/mkl_graph_util.h"
 #include "tensorflow/core/graph/node_builder.h"
 #include "tensorflow/core/lib/core/status.h"
 #include "tensorflow/core/lib/gtl/map_util.h"
@@ -33,8 +36,6 @@ limitations under the License.
 #include "tensorflow/core/platform/logging.h"
 #include "tensorflow/core/util/util.h"
 
-#include "tensorflow/core/graph/mkl_graph_util.h"
-#include "tensorflow/core/graph/mkl_tfconversion_pass.h"
 
 namespace tensorflow {
 

tensorflow/core/common_runtime/mkl_tfconversion_pass.h

@@ -15,8 +15,8 @@ limitations under the License.
 
 // An optimization pass that inserts MklToTf conversion nodes in the graph
 
-#ifndef TENSORFLOW_CORE_GRAPH_MKL_TFCONVERSION_PASS_H_
-#define TENSORFLOW_CORE_GRAPH_MKL_TFCONVERSION_PASS_H_
+#ifndef TENSORFLOW_CORE_COMMON_RUNTIME_MKL_TFCONVERSION_PASS_H_
+#define TENSORFLOW_CORE_COMMON_RUNTIME_MKL_TFCONVERSION_PASS_H_
 
 #ifdef INTEL_MKL
 
@@ -37,4 +37,4 @@ extern bool InsertMklToTfConversionNodes(std::unique_ptr<Graph>* g);
 
 #endif
 
-#endif  // TENSORFLOW_CORE_GRAPH_MKL_TFCONVERSION_PASS_H_
+#endif  // TENSORFLOW_CORE_COMMON_RUNTIME_MKL_TFCONVERSION_PASS_H_

tensorflow/core/common_runtime/mkl_tfconversion_pass_test.cc

@@ -15,16 +15,16 @@ limitations under the License.
 
 #if defined(INTEL_MKL) && defined(ENABLE_MKL)
 
-#include "tensorflow/core/graph/mkl_tfconversion_pass.h"
-#include "tensorflow/core/graph/mkl_graph_util.h"
+#include "tensorflow/core/common_runtime/mkl_tfconversion_pass.h"
 
 #include <algorithm>
 #include <vector>
 
+#include "tensorflow/core/common_runtime/graph_constructor.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/tensor.h"
 #include "tensorflow/core/graph/graph.h"
-#include "tensorflow/core/graph/graph_constructor.h"
+#include "tensorflow/core/graph/mkl_graph_util.h"
 #include "tensorflow/core/graph/testlib.h"
 #include "tensorflow/core/kernels/ops_util.h"
 #include "tensorflow/core/lib/random/simple_philox.h"

tensorflow/core/common_runtime/placer_inspection_required_ops_utils_test.cc

@@ -24,7 +24,6 @@ limitations under the License.
 #include "tensorflow/core/framework/node_def_builder.h"
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/core/common_runtime/placer_test.cc

@@ -25,6 +25,7 @@ limitations under the License.
 #include "tensorflow/core/common_runtime/device_factory.h"
 #include "tensorflow/core/common_runtime/device_set.h"
 #include "tensorflow/core/common_runtime/graph_constructor.h"
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/common_runtime/optimization_registry.h"
 #include "tensorflow/core/framework/device_attributes.pb.h"
 #include "tensorflow/core/framework/function.h"
@@ -36,7 +37,6 @@ limitations under the License.
 #include "tensorflow/core/framework/types.pb.h"
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/errors.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/lib/strings/str_util.h"

tensorflow/core/common_runtime/quantize_training.cc

@@ -13,16 +13,14 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/common_runtime/quantize_training.h"
+
 #include <algorithm>
 #include <atomic>
 #include <set>
 #include <unordered_map>
 #include <vector>
 
-#include "tensorflow/core/graph/quantize_training.h"
-
-#include "tensorflow/core/common_runtime/executor.h"
-#include "tensorflow/core/common_runtime/function.h"
 #include "tensorflow/core/common_runtime/memory_types.h"
 #include "tensorflow/core/framework/log_memory.h"
 #include "tensorflow/core/framework/op_kernel.h"

tensorflow/core/common_runtime/quantize_training.h

@@ -13,8 +13,8 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
-#ifndef TENSORFLOW_CORE_GRAPH_QUANTIZE_TRAINING_H_
-#define TENSORFLOW_CORE_GRAPH_QUANTIZE_TRAINING_H_
+#ifndef TENSORFLOW_CORE_COMMON_RUNTIME_QUANTIZE_TRAINING_H_
+#define TENSORFLOW_CORE_COMMON_RUNTIME_QUANTIZE_TRAINING_H_
 
 #include "tensorflow/core/graph/graph.h"
 
@@ -53,4 +53,4 @@ Status DoQuantizeTrainingOnGraphDef(const GraphDef& input_graphdef,
 
 }  // namespace tensorflow
 
-#endif  // TENSORFLOW_CORE_GRAPH_QUANTIZE_TRAINING_H_
+#endif  // TENSORFLOW_CORE_COMMON_RUNTIME_QUANTIZE_TRAINING_H_

tensorflow/core/common_runtime/quantize_training_test.cc

@@ -13,13 +13,13 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/common_runtime/quantize_training.h"
+
 #include <map>
 #include <string>
 #include <unordered_map>
 #include <vector>
 
-#include "tensorflow/core/graph/quantize_training.h"
-
 #include "tensorflow/core/common_runtime/device_factory.h"
 #include "tensorflow/core/common_runtime/device_mgr.h"
 #include "tensorflow/core/framework/node_def_util.h"

tensorflow/core/graph/BUILD

@@ -18,31 +18,6 @@ cc_library(
     hdrs = ["mkl_graph_util.h"],
 )
 
-# TODO(bmzhao): Refactor this target to use granular dependencies
-# after stage 4 of the TF build refactor is complete:
-# https://github.com/tensorflow/community/pull/179
-tf_cc_test(
-    name = "quantize_training_test",
-    srcs = ["quantize_training_test.cc"],
-    deps = [
-        "//tensorflow/core",
-        "//tensorflow/core:all_kernels",
-        "//tensorflow/core:core_cpu",
-        "//tensorflow/core:core_cpu_internal",
-        "//tensorflow/core:direct_session_internal",
-        "//tensorflow/core:framework",
-        "//tensorflow/core:framework_internal",
-        "//tensorflow/core:lib",
-        "//tensorflow/core:lib_internal",
-        "//tensorflow/core:ops",
-        "//tensorflow/core:protos_all_cc",
-        "//tensorflow/core:test",
-        "//tensorflow/core:test_main",
-        "//tensorflow/core:testlib",
-        "//tensorflow/core/util:protos_test_cc",
-    ],
-)
-
 tf_cc_test(
     name = "collective_order_test",
     size = "small",
@@ -69,11 +44,9 @@ filegroup(
     srcs = [
         "algorithm.h",
         "default_device.h",
-        "gradients.h",
         "graph.h",
         "graph_constructor.h",
         "graph_def_builder.h",
-        "graph_def_builder_util.h",
         "graph_node_util.h",
         "node_builder.h",
         "validate.h",
@@ -110,7 +83,7 @@ filegroup(
 # Note(bmzhao): This target is a holdover from the GRAPH_HDRS array
 # in tensorflow/core/BUILD. This target contains all '.h' files under
 # tensorflow/core/graph, except for the following:
-# 'benchmark_testlib.h', 'mkl_graph_util.h', 'gradients.h', 'quantize_training.h'.
+# 'benchmark_testlib.h'.
 filegroup(
     name = "graph_headers",
     srcs = [
@@ -124,11 +97,8 @@ filegroup(
         "graph.h",
         "graph_constructor.h",  # NOTE(mrry): Don't include the .cc since it depends on common_runtime.
         "graph_def_builder.h",
-        "graph_def_builder_util.h",
         "graph_node_util.h",
         "graph_partition.h",
-        "mkl_layout_pass.h",
-        "mkl_tfconversion_pass.h",
         "node_builder.h",
         "optimizer_cse.h",
         "subgraph.h",
@@ -155,32 +125,6 @@ filegroup(
     ],
 )
 
-filegroup(
-    name = "core_cpu_lib_headers",
-    srcs = [
-        "gradients.h",
-        "quantize_training.h",
-    ],
-)
-
-# This file depends on common_runtime.
-filegroup(
-    name = "core_cpu_base_no_ops_srcs",
-    srcs = [
-        "graph_def_builder_util.cc",
-    ],
-)
-
-filegroup(
-    name = "core_cpu_impl_srcs",
-    srcs = [
-        "gradients.cc",
-        "mkl_layout_pass.cc",
-        "mkl_tfconversion_pass.cc",
-        "quantize_training.cc",
-    ],
-)
-
 filegroup(
     name = "testlib_headers",
     srcs = [
@@ -218,16 +162,6 @@ filegroup(
     ],
 )
 
-filegroup(
-    name = "quantize_training_hdrs",
-    srcs = [
-        "quantize_training.h",
-    ],
-    visibility = [
-        "//tensorflow/python:__pkg__",
-    ],
-)
-
 filegroup(
     name = "mobile_srcs_only_runtime",
     srcs = [
@@ -245,30 +179,20 @@ filegroup(
         "default_device.h",
         "edgeset.cc",
         "edgeset.h",
-        "gradients.cc",
-        "gradients.h",
         "graph.cc",
         "graph.h",
         "graph_constructor.h",
         "graph_def_builder.cc",
         "graph_def_builder.h",
-        "graph_def_builder_util.cc",
-        "graph_def_builder_util.h",
         "graph_node_util.cc",
         "graph_node_util.h",
         "graph_partition.cc",
         "graph_partition.h",
         "mkl_graph_util.h",
-        "mkl_layout_pass.cc",
-        "mkl_layout_pass.h",
-        "mkl_tfconversion_pass.cc",
-        "mkl_tfconversion_pass.h",
         "node_builder.cc",
         "node_builder.h",
         "optimizer_cse.cc",
         "optimizer_cse.h",
-        "quantize_training.cc",
-        "quantize_training.h",
         "subgraph.cc",
         "subgraph.h",
         "tensor_id.cc",

tensorflow/core/graph/algorithm_test.cc

@@ -18,11 +18,11 @@ limitations under the License.
 #include <string>
 #include <vector>
 
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/graph/benchmark_testlib.h"
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/graph/subgraph.h"
 #include "tensorflow/core/kernels/ops_util.h"
 #include "tensorflow/core/lib/core/status.h"

tensorflow/core/graph/collective_order_test.cc

@@ -15,9 +15,9 @@ limitations under the License.
 #include "tensorflow/core/graph/collective_order.h"
 
 #include <gmock/gmock.h>
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/node_def_builder.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"
 

tensorflow/core/graph/graph_def_builder_test.cc

@@ -15,9 +15,9 @@ limitations under the License.
 
 #include "tensorflow/core/graph/graph_def_builder.h"
 
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/versions.pb.h"
 #include "tensorflow/core/graph/graph.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/kernels/ops_util.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/core/graph/subgraph_test.cc

@@ -18,12 +18,12 @@ limitations under the License.
 #include <string>
 #include <vector>
 
+#include "tensorflow/core/common_runtime/graph_def_builder_util.h"
 #include "tensorflow/core/framework/graph.pb.h"
 #include "tensorflow/core/framework/partial_tensor_shape.h"
 #include "tensorflow/core/graph/graph.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/graph/graph_def_builder.h"
-#include "tensorflow/core/graph/graph_def_builder_util.h"
 #include "tensorflow/core/kernels/ops_util.h"
 #include "tensorflow/core/lib/core/status.h"
 #include "tensorflow/core/lib/core/status_test_util.h"

tensorflow/core/kernels/function_ops.cc

@@ -21,11 +21,11 @@ limitations under the License.
 #include "tensorflow/core/common_runtime/device.h"
 #include "tensorflow/core/common_runtime/executor.h"
 #include "tensorflow/core/common_runtime/function.h"
+#include "tensorflow/core/common_runtime/gradients.h"
 #include "tensorflow/core/common_runtime/memory_types.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/register_types.h"
 #include "tensorflow/core/graph/algorithm.h"
-#include "tensorflow/core/graph/gradients.h"
 #include "tensorflow/core/graph/graph_constructor.h"
 #include "tensorflow/core/platform/macros.h"
 #include "tensorflow/core/platform/tracing.h"

tensorflow/core/kernels/quantized_resize_bilinear_op_test.cc

@@ -21,11 +21,11 @@ limitations under the License.
 #include "tensorflow/cc/ops/array_ops.h"
 #include "tensorflow/cc/ops/const_op.h"
 #include "tensorflow/cc/ops/image_ops.h"
+#include "tensorflow/core/common_runtime/gradients.h"
 #include "tensorflow/core/framework/node_def_builder.h"
 #include "tensorflow/core/framework/node_def_util.h"
 #include "tensorflow/core/framework/shape_inference_testutil.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
-#include "tensorflow/core/graph/gradients.h"
 #include "tensorflow/core/kernels/quantization_utils.h"
 #include "tensorflow/core/lib/core/status_test_util.h"
 #include "tensorflow/core/platform/test.h"

tensorflow/python/BUILD

@@ -712,7 +712,7 @@ tf_python_pybind_extension(
     srcs = [
         "training/quantize_training_wrapper.cc",
     ],
-    hdrs = ["//tensorflow/core/graph:quantize_training_hdrs"],
+    hdrs = ["//tensorflow/core/common_runtime:quantize_training_hdrs"],
     module_name = "_pywrap_quantize_training",
     deps = [
         ":pybind11_lib",
@@ -5952,11 +5952,10 @@ filegroup(
         "//tensorflow/c:tf_status_helper",  # tfe
         "//tensorflow/compiler/jit:flags",  #tfe
         "//tensorflow/compiler/mlir/python:mlir",  # mlir
-        "//tensorflow/core/common_runtime:core_cpu_base_no_ops",  # tf_session
-        "//tensorflow/core/common_runtime:core_cpu_rump_impl",  # quantize_training
         "//tensorflow/core/common_runtime:device",  # device_lib, tfe, tf_session
         "//tensorflow/core/common_runtime:device_factory",  # device_lib, tfe, tf_session
         "//tensorflow/core/common_runtime:graph_constructor",  # tf_session
+        "//tensorflow/core/common_runtime:quantize_training",  # quantize_training
         "//tensorflow/core/common_runtime:session_options",  # device_lib, tfe, tf_session
         "//tensorflow/core/common_runtime:session_state",  # tf_session
         "//tensorflow/core/data/service:server_lib",  # server_lib

tensorflow/python/training/quantize_training_wrapper.cc

@@ -14,7 +14,7 @@ limitations under the License.
 ==============================================================================*/
 
 #include "pybind11/pybind11.h"
-#include "tensorflow/core/graph/quantize_training.h"
+#include "tensorflow/core/common_runtime/quantize_training.h"
 #include "tensorflow/python/lib/core/pybind11_lib.h"
 #include "tensorflow/python/lib/core/pybind11_status.h"
 

tensorflow/tools/def_file_filter/symbols_pybind.txt

@@ -86,7 +86,7 @@ tensorflow::DeviceFactory::ListAllPhysicalDevices
 [session_options]  # device_lib, tfe, tf_session
 tensorflow::SessionOptions::SessionOptions
 
-[core_cpu_rump_impl]  # quantize_training
+[quantize_training]  # quantize_training
 tensorflow::DoQuantizeTrainingOnSerializedGraphDef
 
 [session_state]  # tf_session