experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
STT-tensorflow/tensorflow/core/common_runtime
History
Isha Arkatkar 79cd2fd4ac Update check for multiple devices found to check whether there is a single device colocated with the default_device. 
If multiple devices are found in the cluster, but there is a single device colocated with default_device, the single colocated device would be used by placer. This change relaxes the current check to account for parent device scope.

For example,
@tf.function()
def replica_fn()
  with tf.device("gpu:0")

with tf.device("worker1"):
   replica_fn()

In this case there may be multiple gpu:0 in the cluster, but gpu:0 on worker 1 should be picked. In this case, we should not receive multiple devices found error.

PiperOrigin-RevId: 357996270
Change-Id: I6a19deb5f26d7741ba7f7adff7f14550679b3d6d
2021-02-17 11:48:37 -08:00
..
device
Use GpuVirtualMemAllocator in GPUBFCAllocator.
2021-01-29 17:31:11 -08:00
eager
Fix a race on local device manager and rendezvous when setting the collective server on eager context.
2021-02-12 11:12:32 -08:00
gpu
Update benchmarks to newer API
2021-02-09 14:03:25 -08:00
accumulate_n_optimizer.cc
…
allocator_retry.cc
add a metric to track bfc allocator delay.
2020-07-28 14:38:22 -07:00
allocator_retry.h
…
base_collective_executor.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
base_collective_executor.h
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
bfc_allocator_test.cc
Always record memory debug info like op_name and step_id in BFCAllocator Chunk. Remove flag TENSORFLOW_MEM_DEBUG and MEM_DEBUG_SIZE_HISTORY_SIZE because the debug information is initialized by default.
2021-02-04 11:10:01 -08:00
bfc_allocator.cc
This CL introduces a GetDebugOpName function to avoid using op_name if the Chunk is not in use.
2021-02-10 11:57:56 -08:00
bfc_allocator.h
This CL introduces a GetDebugOpName function to avoid using op_name if the Chunk is not in use.
2021-02-10 11:57:56 -08:00
buf_rendezvous_test.cc
Enable BufRendezvous per-entry cancellation.
2020-10-20 13:28:09 -07:00
buf_rendezvous.cc
Deregister cancellation callback before invoking pros_done or cons_done
2020-10-21 16:40:53 -07:00
buf_rendezvous.h
fix typos in core directory
2020-10-29 02:52:55 +03:00
BUILD
Always record memory debug info like op_name and step_id in BFCAllocator Chunk. Remove flag TENSORFLOW_MEM_DEBUG and MEM_DEBUG_SIZE_HISTORY_SIZE because the debug information is initialized by default.
2021-02-04 11:10:01 -08:00
build_graph_options.cc
…
build_graph_options.h
Some const declarations changed to constexpr
2020-04-20 18:19:58 -07:00
collective_executor_mgr_test.cc
Make NcclManager part of CollectiveExecutorMgr
2020-09-17 14:35:16 -07:00
collective_executor_mgr.cc
Make NcclManager part of CollectiveExecutorMgr
2020-09-17 14:35:16 -07:00
collective_executor_mgr.h
Make NcclManager part of CollectiveExecutorMgr
2020-09-17 14:35:16 -07:00
collective_param_resolver_local_test.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
collective_param_resolver_local.cc
Fix a use-after-free of a cancellation manager when a local collective op is cancelled but the group finishes normally
2021-02-10 14:43:43 -08:00
collective_param_resolver_local.h
Allow cancellation of v2 collectives during param resolution
2021-02-05 10:03:57 -08:00
collective_rma_local_test.cc
Set a timeout to check health RPC
2020-10-21 13:02:25 -07:00
collective_rma_local.cc
Set a timeout to check health RPC
2020-10-21 13:02:25 -07:00
collective_rma_local.h
Set a timeout to check health RPC
2020-10-21 13:02:25 -07:00
collective_util.cc
Move devices information from instance params to group params
2020-09-28 17:57:59 -07:00
collective_util.h
…
colocation_graph.cc
Fixed misleading error message
2020-12-30 15:00:32 -08:00
colocation_graph.h
Introduce CompositeDevice which represents a set of physical devices.
2020-04-21 14:44:53 -07:00
composite_device_test.cc
Stop holding custom devices in TensorHandles
2021-01-12 11:22:48 -08:00
composite_device.cc
Stop holding custom devices in TensorHandles
2021-01-12 11:22:48 -08:00
composite_device.h
replace PFLR DeviceGetContext hardcode with Device::IsRemoteCallAllowed
2020-10-22 20:49:03 +02:00
constant_folding_test.cc
[ROCm] Re-enabling unit-tests that are now passing on ROCm platform
2020-12-07 16:41:10 +00:00
constant_folding.cc
Resubmit constant folding change without the 1024byte limit. It was causing tf.where to fail in tf2xla.
2020-10-14 21:51:24 -07:00
constant_folding.h
…
copy_tensor.cc
give sendop/recvop some temporary attribute for tracing. so we can annotate the memcpy device events better.
2020-02-20 10:32:29 -08:00
copy_tensor.h
…
costmodel_manager.cc
…
costmodel_manager.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
debugger_state_interface.cc
…
debugger_state_interface.h
…
device_factory.h
Move Device and DeviceFactory to core/framework. https://github.com/tensorflow/tensorflow/pull/43611/ needs to access device registry to get a subtype for a given device type as a part of kernel registration. Kernel registration is a part of framework and should not depend on common_runtime to avoid a circular dependency.
2020-10-16 15:11:35 -07:00
device_mgr_test.cc
Fix EagerContext::HostCPU().
2020-01-31 11:35:33 -08:00
device_mgr.cc
Validate remote resource devices before safe access of resources.
2020-05-12 21:27:15 -07:00
device_mgr.h
Avoid segmentation fault when instantiating functions on removed devices.
2020-06-11 12:58:56 -07:00
device_resolver_local_test.cc
Use device attributes from group resolution
2020-09-09 10:53:49 -07:00
device_resolver_local.cc
Use device attributes from group resolution
2020-09-09 10:53:49 -07:00
device_resolver_local.h
Use device attributes from group resolution
2020-09-09 10:53:49 -07:00
device_set_test.cc
Refactor sorting logic of PrioritizedDeviceTypeVector.
2020-01-31 10:16:53 -08:00
device_set.cc
Take device locality into account during prioritization.
2020-05-21 12:22:18 -07:00
device_set.h
Take device locality into account during prioritization.
2020-05-21 12:22:18 -07:00
device.h
Move Device and DeviceFactory to core/framework. https://github.com/tensorflow/tensorflow/pull/43611/ needs to access device registry to get a subtype for a given device type as a part of kernel registration. Kernel registration is a part of framework and should not depend on common_runtime to avoid a circular dependency.
2020-10-16 15:11:35 -07:00
direct_session_test.cc
Internal tests cleanup.
2020-10-29 23:45:48 -07:00
direct_session_with_tracking_alloc_test.cc
Delete SYCL support
2020-09-15 11:21:47 -07:00
direct_session.cc
Internal change
2020-12-14 18:08:03 -08:00
direct_session.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
dma_helper.h
…
dynamic_device_mgr_test.cc
Avoid segmentation fault when instantiating functions on removed devices.
2020-06-11 12:58:56 -07:00
dynamic_device_mgr.cc
Avoid segmentation fault when instantiating functions on removed devices.
2020-06-11 12:58:56 -07:00
entry.h
[Executor] Split ExecutorState into PropagatorState and ExecutorState<PropagatorStateType>.
2020-04-01 14:35:21 -07:00
eval_const_tensor.cc
Support constant folding across Arg nodes during shape inference on nested function calls.
2020-10-19 21:27:26 -07:00
eval_const_tensor.h
Support constant folding across Arg nodes during shape inference on nested function calls.
2020-10-19 21:27:26 -07:00
executor_factory.cc
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
executor_factory.h
…
executor_test.cc
Internal tests cleanup.
2020-10-29 23:45:48 -07:00
executor.cc
Internal change
2020-12-14 18:08:03 -08:00
executor.h
Internal change
2020-12-10 21:07:03 -08:00
function_body.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
function_body.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
function_def_utils.cc
Don't check for if null after already dereferenced
2020-12-30 14:42:46 -08:00
function_def_utils.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
function_optimization_registration_test.cc
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_optimization_registry_no_pass_test.cc
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_optimization_registry_pass_failure_test.cc
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_optimization_registry_test.cc
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_optimization_registry.cc
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_optimization_registry.h
Add separate registry for function graph optimization passes in function runtime.
2020-02-06 15:37:07 -08:00
function_test.cc
Function/control flow inlining: only use the function's job/replica/task placement if the body node doesn't have its own
2020-12-15 16:59:01 -08:00
function_testlib.cc
…
function_testlib.h
…
function_threadpool_test.cc
[TF2XLA] Remove the serialization of CustomKernelCreator, since there is only one, and we won't add new ones
2020-09-22 10:59:59 -07:00
function_utils.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
function_utils.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
function.cc
[TF2XLA] Remove the serialization of CustomKernelCreator, since there is only one, and we won't add new ones
2020-09-22 10:59:59 -07:00
function.h
[TF2XLA] Remove the serialization of CustomKernelCreator, since there is only one, and we won't add new ones
2020-09-22 10:59:59 -07:00
gpu_device_context.h
…
gradients.cc
tensorflow/core resolutions set 1
2020-07-26 23:49:59 +00:00
gradients.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
graph_constructor_test.cc
[Build cleanup] Remove "tensorflow/core/graph/graph_constructor.h".
2020-05-04 10:20:25 -07:00
graph_constructor.cc
Add minor optimization for graph copies.
2021-02-01 16:55:28 -08:00
graph_constructor.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
graph_def_builder_util.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
graph_def_builder_util.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
graph_execution_state.cc
Set collective_graph_key when there're v2 collective ops present
2021-01-27 11:36:14 -08:00
graph_execution_state.h
Make GraphExecutionState::OptimizeGraph() a public method so that it can reused by other components.
2021-02-01 10:26:25 -08:00
graph_optimizer.cc
[TF/XLA] Rollback of rollback of 313256383, with a UB fix.
2020-05-26 20:47:42 -07:00
graph_optimizer.h
[TF/XLA] Rollback of rollback of 313256383, with a UB fix.
2020-05-26 20:47:42 -07:00
graph_runner_test.cc
Clean up tensor_testutil.
2020-07-02 11:34:23 -07:00
graph_runner.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
graph_runner.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (3/n).
2020-04-27 10:43:23 -07:00
graph_view.cc
[Executor] Avoid unnecessary NodeItem::input_type() calls in PrepareInputs().
2020-05-08 09:41:13 -07:00
graph_view.h
[Executor] Avoid unnecessary NodeItem::input_type() calls in PrepareInputs().
2020-05-08 09:41:13 -07:00
hierarchical_tree_broadcaster_test.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
hierarchical_tree_broadcaster.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
hierarchical_tree_broadcaster.h
Share ownership of CollectiveContext with kernels.
2020-07-15 13:32:36 -07:00
immutable_executor_state.cc
[Executor] Optimize PropagatorState::FindOrCreateChildFrame().
2020-05-07 17:26:18 -07:00
immutable_executor_state.h
[Executor] Optimize PropagatorState::FindOrCreateChildFrame().
2020-05-07 17:26:18 -07:00
inline_function_utils.cc
Function/control flow inlining: only use the function's job/replica/task placement if the body node doesn't have its own
2020-12-15 16:59:01 -08:00
inline_function_utils.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
input_colocation_exemption_registry_test.cc
…
input_colocation_exemption_registry.cc
…
input_colocation_exemption_registry.h
…
inspecting_placer.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
inspecting_placer.h
…
isolate_placer_inspection_required_ops_pass_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
isolate_placer_inspection_required_ops_pass.cc
…
isolate_placer_inspection_required_ops_pass.h
…
kernel_benchmark_testlib.cc
internal framework cleanup
2020-10-24 11:27:07 -07:00
kernel_benchmark_testlib.h
Deprecated old API
2020-11-23 22:47:15 -08:00
local_device.cc
…
local_device.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
local_executor_params.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (2/n).
2020-04-22 15:36:40 -07:00
lower_case_op_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
lower_case_op.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_case_op.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_function_call_op_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
lower_function_call_op.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_function_call_op.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_functional_ops_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
lower_functional_ops.cc
generate stateless_case op if all ops in all branches are stateless
2020-07-30 16:23:47 -07:00
lower_functional_ops.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_if_op_test.cc
Rollback changelist 338246477
2020-12-09 09:23:02 -08:00
lower_if_op.cc
Rollback changelist 338246477
2020-12-09 09:23:02 -08:00
lower_if_op.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
lower_while_op_test.cc
Fix requested_device of NextIteration node.
2020-05-08 14:33:01 -07:00
lower_while_op.cc
Fix requested_device of NextIteration node.
2020-05-08 14:33:01 -07:00
lower_while_op.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
memory_types_test.cc
Delete SYCL support
2020-09-15 11:21:47 -07:00
memory_types.cc
Delete SYCL support
2020-09-15 11:21:47 -07:00
memory_types.h
…
metrics.h
[tf.data] Adding a metric for bytes produced and consumed by individual transformations, refactoring infrastructure for recording tf.data metrics, and moving the metrics API and implementation from common_runtime to framework.
2020-04-06 10:39:03 -07:00
mkl_cpu_allocator_test.cc
…
mkl_cpu_allocator.cc
Removed MKL blob for Window/Mac and fixed openMP build for Windows.
2020-10-17 16:33:33 -07:00
mkl_cpu_allocator.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
mkl_layout_pass_test.cc
Merge pull request #45115 from Intel-tensorflow:dnn0x_clean_runtime
2021-01-05 09:15:26 -08:00
mkl_layout_pass.cc
Merge pull request #45115 from Intel-tensorflow:dnn0x_clean_runtime
2021-01-05 09:15:26 -08:00
mkl_layout_pass.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
mkl_tfconversion_pass_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
mkl_tfconversion_pass.cc
Changes to graph rewrite to support native format
2020-08-26 20:01:22 -07:00
mkl_tfconversion_pass.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
mkl_threadpool_device_test.cc
Merge pull request #43795 from Intel-tensorflow:vishakh1/openmp_changes
2020-10-16 13:22:42 -07:00
optimization_registry_test.cc
…
optimization_registry.cc
track total time besides per pass time for KPI consumption.
2020-07-27 12:49:09 -07:00
optimization_registry.h
minor spelling tweaks
2020-02-27 15:42:16 +09:00
parallel_concat_optimizer.cc
Rolling forward "[Build cleanup] Split "core_cpu_impl" into fine-grained targets (1/n)."
2020-04-22 09:43:00 -07:00
partitioning_utils_test.cc
Support packed tensor inputs in ProcessFunctionLibraryRuntime.
2020-05-04 11:18:39 -07:00
partitioning_utils.cc
NFC: Refactored existing runtime code to prepare for TFRT/TF2 training's integration points.
2020-12-15 13:26:34 -08:00
partitioning_utils.h
NFC: Refactored existing runtime code to prepare for TFRT/TF2 training's integration points.
2020-12-15 13:26:34 -08:00
pending_counts_test.cc
This follows a similar approach to the earlier work by Derek Murray
2020-09-15 17:15:11 -07:00
pending_counts.h
Fix CHECK on alignment of PendingCounts storage
2020-09-23 14:15:49 -07:00
permuter_test.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
permuter.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
permuter.h
Atomically update status_ in permuter.
2020-08-19 16:24:36 -07:00
placer_inspection_required_ops_utils_test.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
placer_inspection_required_ops_utils.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (5/5).
2020-04-28 20:24:52 -07:00
placer_inspection_required_ops_utils.h
…
placer_test.cc
Fixed misleading error message
2020-12-30 15:00:32 -08:00
placer.cc
Split out node_def_util, op_def_builder, op_def_util and attr_value_util targets in tensorflow/core/framework/BUILD. Split out
2020-01-06 11:46:01 -08:00
placer.h
…
pool_allocator.cc
Enable SubAllocator to inform the main allocator of an over-allocation.
2020-11-20 17:56:43 -08:00
pool_allocator.h
Support coalescing of adjacent allocation regions in BFC allocator.
2021-01-13 15:47:59 -08:00
process_function_library_runtime_test.cc
[TF2XLA] Remove the serialization of CustomKernelCreator, since there is only one, and we won't add new ones
2020-09-22 10:59:59 -07:00
process_function_library_runtime.cc
Update check for multiple devices found to check whether there is a single device colocated with the default_device.
2021-02-17 11:48:37 -08:00
process_function_library_runtime.h
Refresh device in EagerContext and pflr when device is updated. This is required to allow RuntimeFallback and KernelFallback to access TPU device created by tfrt.
2021-02-05 14:52:22 -08:00
process_state.cc
Fix some comment syntax; NFC
2020-12-23 15:18:36 -08:00
process_state.h
Remove mutex lock from the ProcessState::GetCPUAllocator.
2020-07-19 18:48:15 -07:00
process_util_test.cc
…
process_util.cc
minor cleanup.
2020-04-16 12:08:58 -07:00
process_util.h
…
profile_handler.h
…
propagator_debug_utils.cc
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (2/n).
2020-04-22 15:36:40 -07:00
propagator_debug_utils.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (2/n).
2020-04-22 15:36:40 -07:00
propagator_state.cc
This follows a similar approach to the earlier work by Derek Murray
2020-09-15 17:15:11 -07:00
propagator_state.h
This follows a similar approach to the earlier work by Derek Murray
2020-09-15 17:15:11 -07:00
quantize_training_test.cc
[Build cleanup] Update #includes of moved header "graph/graph_constructor.h".
2020-04-29 09:20:48 -07:00
quantize_training.cc
[Build cleanup] Update #includes of moved header "graph/graph_constructor.h".
2020-04-29 09:20:48 -07:00
quantize_training.h
[Build cleanup] Split "core_cpu_impl" into fine-grained targets (4/n).
2020-04-28 09:53:59 -07:00
renamed_device.cc
…
renamed_device.h
Merge pull request #43812 from ppiskorski:get_device_context
2020-10-23 18:14:22 -07:00
rendezvous_mgr.cc
Add Ref/Unref to make sure rendezvous outlives its cancellation callback.
2020-10-02 11:23:42 -07:00
rendezvous_mgr.h
…
rendezvous_util_test.cc
…
rendezvous_util.cc
…
rendezvous_util.h
…
replicate_per_replica_nodes_test.cc
Remove dead replicated Arg nodes.
2020-08-16 16:02:28 -07:00
replicate_per_replica_nodes.cc
Remove dead replicated Arg nodes.
2020-08-16 16:02:28 -07:00
replicate_per_replica_nodes.h
Support packed tensor inputs in ProcessFunctionLibraryRuntime.
2020-05-04 11:18:39 -07:00
ring_alg.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
ring_alg.h
Share ownership of CollectiveContext with kernels.
2020-07-15 13:32:36 -07:00
ring_gatherer_test.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
ring_gatherer.cc
Move devices information from instance params to group params
2020-09-28 17:57:59 -07:00
ring_gatherer.h
…
ring_reducer_test.cc
Ensure that CollectiveParams outlives all references to it.
2021-02-04 10:07:20 -08:00
ring_reducer.cc
Make V2 collective ComputeAsync thread safe
2020-10-21 11:17:35 -07:00
ring_reducer.h
…
scoped_allocator_mgr_test.cc
…
scoped_allocator_mgr.cc
fix typos in core directory
2020-10-29 02:52:55 +03:00
scoped_allocator_mgr.h
Qualify uses of std::string
2020-08-01 12:23:19 -07:00
scoped_allocator.cc
…
scoped_allocator.h
Qualify uses of std::string
2020-08-01 12:23:19 -07:00
session_factory.cc
…
session_factory.h
…
session_options.cc
…
session_state.cc
Export the TF Session classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.
2020-01-29 18:29:05 -08:00
session_test.cc
…
session.cc
Avoid crash in C API and log error.
2020-06-03 13:30:21 -07:00
shape_refiner_test.cc
Clearly indicate when to disable optimizations
2020-04-30 10:37:25 -07:00
shape_refiner.cc
Support constant folding across Arg nodes during shape inference on nested function calls.
2020-10-19 21:27:26 -07:00
shape_refiner.h
Support constant folding across Arg nodes during shape inference on nested function calls.
2020-10-19 21:27:26 -07:00
shared_counter.h
…
simple_propagator_state.cc
[Executor] Fix segfault when using verbose logging in the executor module.
2020-05-08 16:11:54 -07:00
simple_propagator_state.h
[Executor] Fix segfault when using verbose logging in the executor module.
2020-05-08 16:11:54 -07:00
single_threaded_cpu_device.cc
…
single_threaded_cpu_device.h
…
stats_publisher_interface.cc
…
stats_publisher_interface.h
…
step_stats_collector.cc
Optimize calls to std::string::find() and friends for a single char.
2020-12-17 18:11:13 -08:00
step_stats_collector.h
Rolling forward "[Build cleanup] Split "core_cpu_impl" into fine-grained targets (1/n)."
2020-04-22 09:43:00 -07:00
test_collective_executor_mgr.h
Add NcclManager::StartAbort
2020-09-18 13:27:17 -07:00
testlib_ops.cc
…
threadpool_device_factory.cc
…
threadpool_device_test.cc
…
threadpool_device.cc
Merge pull request #43795 from Intel-tensorflow:vishakh1/openmp_changes
2020-10-16 13:22:42 -07:00
threadpool_device.h
…