Struct tensorflow_proto::tensorflow::gpu_options::Experimental

pub struct Experimental {
    pub virtual_devices: Vec<VirtualDevices>,
    pub use_unified_memory: bool,
    pub num_dev_to_dev_copy_streams: i32,
    pub collective_ring_order: String,
    pub timestamped_allocator: bool,
    pub kernel_tracker_max_interval: i32,
    pub kernel_tracker_max_bytes: i32,
    pub kernel_tracker_max_pending: i32,
}

Fields

virtual_devices: Vec<VirtualDevices>

The multi virtual device settings. If empty (not set), it will create single virtual device on each visible GPU, according to the settings in "visible_device_list" above. Otherwise, the number of elements in the list must be the same as the number of visible GPUs (after "visible_device_list" filtering if it is set), and the string represented device names (e.g. /device:GPU:) will refer to the virtual devices and have the field assigned sequentially starting from 0, according to the order they appear in this list and the "memory_limit" list inside each element. For example, visible_device_list = "1,0" virtual_devices { memory_limit: 1GB memory_limit: 2GB } virtual_devices {} will create three virtual devices as: /device:GPU:0 -> visible GPU 1 with 1GB memory /device:GPU:1 -> visible GPU 1 with 2GB memory /device:GPU:2 -> visible GPU 0 with all available memory

NOTE:

1. It's invalid to set both this and "per_process_gpu_memory_fraction" at the same time.
2. Currently this setting is per-process, not per-session. Using different settings in different sessions within same process will result in undefined behavior.

use_unified_memory: bool

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.

num_dev_to_dev_copy_streams: i32

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.

collective_ring_order: String

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.

timestamped_allocator: bool

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.

kernel_tracker_max_interval: i32

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.

kernel_tracker_max_bytes: i32

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.

kernel_tracker_max_pending: i32

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.

Trait Implementations

impl Clone for Experimental

pub fn clone(&self) -> Experimental

Returns a copy of the value.

pub fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Experimental

pub fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Experimental

pub fn default() -> Self

Returns the "default value" for a type.

impl Message for Experimental

pub fn encode_raw<B>(&self, buf: &mut B) where
    B: BufMut, 

pub fn merge_field<B>(
    &mut self,
    tag: u32,
    wire_type: WireType,
    buf: &mut B,
    ctx: DecodeContext
) -> Result<(), DecodeError> where
    B: Buf, 

pub fn encoded_len(&self) -> usize

Returns the encoded length of the message without a length delimiter.

pub fn clear(&mut self)

Clears the message, resetting all fields to their default.

pub fn encode<B>(&self, buf: &mut B) -> Result<(), EncodeError> where
    B: BufMut, 

Encodes the message to a buffer.

pub fn encode_length_delimited<B>(&self, buf: &mut B) -> Result<(), EncodeError> where
    B: BufMut, 

Encodes the message with a length-delimiter to a buffer.

pub fn decode<B>(buf: B) -> Result<Self, DecodeError> where
    Self: Default,
    B: Buf, 

Decodes an instance of the message from a buffer.

pub fn decode_length_delimited<B>(buf: B) -> Result<Self, DecodeError> where
    Self: Default,
    B: Buf, 

Decodes a length-delimited instance of the message from the buffer.

pub fn merge<B>(&mut self, buf: B) -> Result<(), DecodeError> where
    B: Buf, 

Decodes an instance of the message from a buffer, and merges it into self.

pub fn merge_length_delimited<B>(&mut self, buf: B) -> Result<(), DecodeError> where
    B: Buf, 

Decodes a length-delimited instance of the message from buffer, and merges it into self.

impl PartialEq<Experimental> for Experimental

pub fn eq(&self, other: &Experimental) -> bool

This method tests for self and other values to be equal, and is used by ==.

pub fn ne(&self, other: &Experimental) -> bool

This method tests for !=.

impl StructuralPartialEq for Experimental