org.tensorflow.framework

Interface ConfigProto.ExperimentalOrBuilder

All Superinterfaces:

com.google.protobuf.MessageLiteOrBuilder, com.google.protobuf.MessageOrBuilder

All Known Implementing Classes:

ConfigProto.Experimental, ConfigProto.Experimental.Builder

Enclosing class:

ConfigProto

public static interface ConfigProto.ExperimentalOrBuilder
extends com.google.protobuf.MessageOrBuilder

Method Summary

Modifier and Type	Method and Description
boolean	getCollectiveDeterministicSequentialExecution() If true, make collective op execution order sequential and deterministic for potentially concurrent collective instances.
String	getCollectiveGroupLeader() Task name for group resolution.
com.google.protobuf.ByteString	getCollectiveGroupLeaderBytes() Task name for group resolution.
boolean	getCollectiveNccl() If true, use NCCL for CollectiveOps.
boolean	getDisableThreadSpinning() If using a direct session, disable spinning while waiting for work in the thread pool.
String	getExecutorType() Which executor to use, the default executor will be used if it is an empty string or "DEFAULT"
com.google.protobuf.ByteString	getExecutorTypeBytes() Which executor to use, the default executor will be used if it is an empty string or "DEFAULT"
int	getRecvBufMaxChunk() Guidance to formatting of large RecvBuf fields for transfer.
boolean	getShareClusterDevicesInSession() When true, WorkerSessions are created with device attributes from the full cluster.
boolean	getShareSessionStateInClusterspecPropagation() In the following, session state means the value of a variable, elements in a hash table, or any other resource, accessible by worker sessions held by a TF server.
boolean	getUseNumaAffinity() If true, and supported by the platform, the runtime will attempt to use NUMA affinity where applicable.

Methods inherited from interface com.google.protobuf.MessageOrBuilder

findInitializationErrors, getAllFields, getDefaultInstanceForType, getDescriptorForType, getField, getInitializationErrorString, getOneofFieldDescriptor, getRepeatedField, getRepeatedFieldCount, getUnknownFields, hasField, hasOneof

Methods inherited from interface com.google.protobuf.MessageLiteOrBuilder

isInitialized

Method Detail

getCollectiveGroupLeader

String getCollectiveGroupLeader()

 Task name for group resolution.
 

string collective_group_leader = 1;

getCollectiveGroupLeaderBytes

com.google.protobuf.ByteString getCollectiveGroupLeaderBytes()

 Task name for group resolution.
 

string collective_group_leader = 1;

getExecutorType

String getExecutorType()

 Which executor to use, the default executor will be used
 if it is an empty string or "DEFAULT"
 

string executor_type = 3;

getExecutorTypeBytes

com.google.protobuf.ByteString getExecutorTypeBytes()

 Which executor to use, the default executor will be used
 if it is an empty string or "DEFAULT"
 

string executor_type = 3;

getRecvBufMaxChunk

int getRecvBufMaxChunk()

 Guidance to formatting of large RecvBuf fields for transfer.
 Any positive value sets the max chunk size.  0 defaults to 4096.
 Any negative value indicates no max, i.e. one chunk only.
 

int32 recv_buf_max_chunk = 4;

getUseNumaAffinity

boolean getUseNumaAffinity()

 If true, and supported by the platform, the runtime will attempt to
 use NUMA affinity where applicable.  One consequence will be the
 existence of as many CPU devices as there are available NUMA nodes.
 

bool use_numa_affinity = 5;

getCollectiveDeterministicSequentialExecution

boolean getCollectiveDeterministicSequentialExecution()

 If true, make collective op execution order sequential and deterministic
 for potentially concurrent collective instances.
 

bool collective_deterministic_sequential_execution = 6;

getCollectiveNccl

boolean getCollectiveNccl()

 If true, use NCCL for CollectiveOps.  This feature is highly
 experimental.
 

bool collective_nccl = 7;

getShareSessionStateInClusterspecPropagation

boolean getShareSessionStateInClusterspecPropagation()

 In the following, session state means the value of a variable, elements
 in a hash table, or any other resource, accessible by worker sessions
 held by a TF server.
 When ClusterSpec propagation is enabled, the value of
 isolate_session_state is ignored when deciding whether to share session
 states in a TF server (for backwards compatibility reasons).
 - If share_session_state_in_clusterspec_propagation is true, the session
 states are shared.
 - If share_session_state_in_clusterspec_propagation is false, session
 states are isolated.
 When clusterspec propagation is not used, the value of
 share_session_state_in_clusterspec_propagation is ignored when deciding
 whether to share session states in a TF server.
 - If isolate_session_state is true, session states are isolated.
 - If isolate_session_state is false, session states are shared.
 TODO(b/129330037): Add a single API that consistently treats
 isolate_session_state and ClusterSpec propagation.
 

bool share_session_state_in_clusterspec_propagation = 8;

getDisableThreadSpinning

boolean getDisableThreadSpinning()

 If using a direct session, disable spinning while waiting for work in
 the thread pool. This may result in higher latency for completing ops,
 but in the case where there is a lot of spinning may result in lower
 CPU usage.
 

bool disable_thread_spinning = 9;

getShareClusterDevicesInSession

boolean getShareClusterDevicesInSession()

 When true, WorkerSessions are created with device attributes from the
 full cluster.
 This is helpful when a worker wants to partition a graph
 (for example during a PartitionedCallOp).
 

bool share_cluster_devices_in_session = 10;