Esta página foi traduzida pela API Cloud Translation.

GPUOptions.Experimental

classe final estática pública GPUOptions.Experimental

Tipo de protobuf tensorflow.GPUOptions.Experimental

Classes aninhadas

aula	GPUOptions.Experimental.Builder	Tipo de protobuf `tensorflow.GPUOptions.Experimental`
aula	GPUOptions.Experimental.VirtualDevices	Configuration for breaking down a visible GPU into multiple "virtual" devices.
interface	GPUOptions.Experimental.VirtualDevicesOrBuilder

Constantes

interno	COLLECTIVE_RING_ORDER_FIELD_NUMBER
interno	KERNEL_TRACKER_MAX_BYTES_FIELD_NUMBER
interno	KERNEL_TRACKER_MAX_INTERVAL_FIELD_NUMBER
interno	KERNEL_TRACKER_MAX_PENDING_FIELD_NUMBER
interno	NUM_DEV_TO_DEV_COPY_STREAMS_FIELD_NUMBER
interno	TIMESTAMPED_ALLOCATOR_FIELD_NUMBER
interno	USE_UNIFIED_MEMORY_FIELD_NUMBER
interno	VIRTUAL_DEVICES_FIELD_NUMBER

Métodos Públicos

booleano	é igual (objeto obj)
Corda	getCollectiveRingOrder () If non-empty, defines a good GPU ring order on a single worker based on device interconnect.
com.google.protobuf.ByteString	getCollectiveRingOrderBytes () If non-empty, defines a good GPU ring order on a single worker based on device interconnect.
GPUOptions estáticas.Experimental	getDefaultInstance ()
GPUOptions.Experimental	getDefaultInstanceForType ()
final estático com.google.protobuf.Descriptors.Descriptor	getDescriptor ()
interno	getKernelTrackerMaxBytes () If kernel_tracker_max_bytes = n > 0, then a tracking event is inserted after every series of kernels allocating a sum of memory >= n.
interno	getKernelTrackerMaxInterval () Parameters for GPUKernelTracker.
interno	getKernelTrackerMaxPending () If kernel_tracker_max_pending > 0 then no more than this many tracking events can be outstanding at a time.
interno	getNumDevToDevCopyStreams () If > 1, the number of device-to-device copy streams to create for each GPUDevice.
	getParserForType ()
interno	getSerializedSize ()
booleano	getTimestampedAllocator () If true then extra work is done by GPUDevice and GPUBFCAllocator to keep track of when GPU memory is freed and when kernels actually complete so that we can know when a nominally free memory chunk is really not subject to pending use.
final com.google.protobuf.UnknownFieldSet	getUnknownFields ()
booleano	getUseUnifiedMemory () If true, uses CUDA unified memory for memory allocations.
GPUOptions.Experimental.VirtualDevices	getVirtualDevices (índice interno) The multi virtual device settings.
interno	getVirtualDevicesCount () The multi virtual device settings.
Lista< GPUOptions.Experimental.VirtualDevices >	getVirtualDevicesList () The multi virtual device settings.
GPUOptions.Experimental.VirtualDevicesOrBuilder	getVirtualDevicesOrBuilder (índice interno) The multi virtual device settings.
Lista<? estende GPUOptions.Experimental.VirtualDevicesOrBuilder >	getVirtualDevicesOrBuilderList () The multi virtual device settings.
interno	código hash ()
booleano final	éinicializado ()
GPUOptions.Experimental.Builder estático	newBuilder ( GPUOptions.Protótipo experimental)
GPUOptions.Experimental.Builder estático	novoConstrutor ()
GPUOptions.Experimental.Builder	novoBuilderForType ()
GPUOptions estáticas.Experimental	parseDelimitedFrom (entrada InputStream)
GPUOptions estáticas.Experimental	parseDelimitedFrom (entrada InputStream, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
GPUOptions estáticas.Experimental	parseFrom (dados de ByteBuffer)
GPUOptions estáticas.Experimental	parseFrom (entrada com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
GPUOptions estáticas.Experimental	parseFrom (dados de ByteBuffer, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
GPUOptions estáticas.Experimental	parseFrom (entrada com.google.protobuf.CodedInputStream)
GPUOptions estáticas.Experimental	parseFrom (byte[] dados, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
GPUOptions estáticas.Experimental	parseFrom (dados com.google.protobuf.ByteString)
GPUOptions estáticas.Experimental	parseFrom (entrada InputStream, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
GPUOptions estáticas.Experimental	parseFrom (dados com.google.protobuf.ByteString, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
estático	analisador ()
GPUOptions.Experimental.Builder	paraConstrutor ()
vazio	writeTo (saída com.google.protobuf.CodedOutputStream)

Métodos herdados

Da classe java.lang.Object

booleano	é igual (objeto arg0)
aula final<?>	getClass ()
interno	código hash ()
vazio final	notificar ()
vazio final	notificar todos ()
Corda	toString ()
vazio final	espere (long arg0, int arg1)
vazio final	espere (arg0 longo)
vazio final	espere ()

Da interface org.tensorflow.proto.framework.GPUOptions.ExperimentalOrBuilder

cadeia abstrata	getCollectiveRingOrder () If non-empty, defines a good GPU ring order on a single worker based on device interconnect.
abstrato com.google.protobuf.ByteString	getCollectiveRingOrderBytes () If non-empty, defines a good GPU ring order on a single worker based on device interconnect.
abstrato int	getKernelTrackerMaxBytes () If kernel_tracker_max_bytes = n > 0, then a tracking event is inserted after every series of kernels allocating a sum of memory >= n.
abstrato int	getKernelTrackerMaxInterval () Parameters for GPUKernelTracker.
abstrato int	getKernelTrackerMaxPending () If kernel_tracker_max_pending > 0 then no more than this many tracking events can be outstanding at a time.
abstrato int	getNumDevToDevCopyStreams () If > 1, the number of device-to-device copy streams to create for each GPUDevice.
booleano abstrato	getTimestampedAllocator () If true then extra work is done by GPUDevice and GPUBFCAllocator to keep track of when GPU memory is freed and when kernels actually complete so that we can know when a nominally free memory chunk is really not subject to pending use.
booleano abstrato	getUseUnifiedMemory () If true, uses CUDA unified memory for memory allocations.
GPUOptions.Experimental.VirtualDevices abstrato	getVirtualDevices (índice interno) The multi virtual device settings.
abstrato int	getVirtualDevicesCount () The multi virtual device settings.
Lista abstrata< GPUOptions.Experimental.VirtualDevices >	getVirtualDevicesList () The multi virtual device settings.
abstrato GPUOptions.Experimental.VirtualDevicesOrBuilder	getVirtualDevicesOrBuilder (índice interno) The multi virtual device settings.
lista abstrata<? estende GPUOptions.Experimental.VirtualDevicesOrBuilder >	getVirtualDevicesOrBuilderList () The multi virtual device settings.

Constantes

público estático final int COLLECTIVE_RING_ORDER_FIELD_NUMBER

Valor Constante: 4

público estático final int KERNEL_TRACKER_MAX_BYTES_FIELD_NUMBER

Valor Constante: 8

público estático final int KERNEL_TRACKER_MAX_INTERVAL_FIELD_NUMBER

Valor Constante: 7

público estático final int KERNEL_TRACKER_MAX_PENDING_FIELD_NUMBER

Valor Constante: 9

público estático final int NUM_DEV_TO_DEV_COPY_STREAMS_FIELD_NUMBER

Valor Constante: 3

público estático final int TIMESTAMPED_ALLOCATOR_FIELD_NUMBER

Valor Constante: 5

público estático final int USE_UNIFIED_MEMORY_FIELD_NUMBER

Valor Constante: 2

público estático final int VIRTUAL_DEVICES_FIELD_NUMBER

Valor Constante: 1

Métodos Públicos

booleano público é igual (Object obj)

String pública getCollectiveRingOrder ()

 If non-empty, defines a good GPU ring order on a single worker based on
 device interconnect.  This assumes that all workers have the same GPU
 topology.  Specify as a comma-separated string, e.g. "3,2,1,0,7,6,5,4".
 This ring order is used by the RingReducer implementation of
 CollectiveReduce, and serves as an override to automatic ring order
 generation in OrderTaskDeviceMap() during CollectiveParam resolution.

string collective_ring_order = 4;

público com.google.protobuf.ByteString getCollectiveRingOrderBytes ()

 If non-empty, defines a good GPU ring order on a single worker based on
 device interconnect.  This assumes that all workers have the same GPU
 topology.  Specify as a comma-separated string, e.g. "3,2,1,0,7,6,5,4".
 This ring order is used by the RingReducer implementation of
 CollectiveReduce, and serves as an override to automatic ring order
 generation in OrderTaskDeviceMap() during CollectiveParam resolution.

string collective_ring_order = 4;

GPUOptions estática pública.Experimental getDefaultInstance ()

público GPUOptions.Experimental getDefaultInstanceForType ()

final estático público com.google.protobuf.Descriptors.Descriptor getDescriptor ()

público int getKernelTrackerMaxBytes ()

 If kernel_tracker_max_bytes = n > 0, then a tracking event is
 inserted after every series of kernels allocating a sum of
 memory >= n.  If one kernel allocates b * n bytes, then one
 event will be inserted after it, but it will count as b against
 the pending limit.

int32 kernel_tracker_max_bytes = 8;

público int getKernelTrackerMaxInterval ()

 Parameters for GPUKernelTracker.  By default no kernel tracking is done.
 Note that timestamped_allocator is only effective if some tracking is
 specified.
 If kernel_tracker_max_interval = n > 0, then a tracking event
 is inserted after every n kernels without an event.

int32 kernel_tracker_max_interval = 7;

público int getKernelTrackerMaxPending ()

 If kernel_tracker_max_pending > 0 then no more than this many
 tracking events can be outstanding at a time.  An attempt to
 launch an additional kernel will stall until an event
 completes.

int32 kernel_tracker_max_pending = 9;

público int getNumDevToDevCopyStreams ()

 If > 1, the number of device-to-device copy streams to create
 for each GPUDevice.  Default value is 0, which is automatically
 converted to 1.

int32 num_dev_to_dev_copy_streams = 3;

público getParserForType ()

público int getSerializedSize ()

getTimestampedAllocator booleano público ()

 If true then extra work is done by GPUDevice and GPUBFCAllocator to
 keep track of when GPU memory is freed and when kernels actually
 complete so that we can know when a nominally free memory chunk
 is really not subject to pending use.

bool timestamped_allocator = 5;

final público com.google.protobuf.UnknownFieldSet getUnknownFields ()

getUseUnifiedMemory booleano público ()

 If true, uses CUDA unified memory for memory allocations. If
 per_process_gpu_memory_fraction option is greater than 1.0, then unified
 memory is used regardless of the value for this field. See comments for
 per_process_gpu_memory_fraction field for more details and requirements
 of the unified memory. This option is useful to oversubscribe memory if
 multiple processes are sharing a single GPU while individually using less
 than 1.0 per process memory fraction.

bool use_unified_memory = 2;