Tensorflow 跨设备通信

__Sunny__

于 2017-06-01 15:45:14 发布

阅读量2.3k

点赞数

分类专栏： TensorFlow 文章标签： TensorFlow

本文链接：https://blog.csdn.net/s_sunnyy/article/details/72830935

版权

TensorFlow 使用 Send 和 Recv 操作实现跨设备通信，当图在不同设备间分割时自动添加。Send op 立即返回，而 Recv op 注册回调以在张量可用时产生输出并解除阻塞后续计算。数据传输通过 DMA 例程在相同进程的设备间进行，远程传输则由 GrpcRemoteRendezvous 类使用 gRPC 处理。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

参考 Tensorflow Cross Device Communication

In TensorFlow, cross-device communication is achieved using the Rendezvous interface, which has multiple different implementations, depending on the deployment. The comment on that interface describes the general idea:

在 TensorFlow 中，跨设备通信使用Rendezvous接口实现，该接口具有多个不同的实现，具体取决于不同的部署。该接口的注释描述了大概的思路：

rendezvous.h

// A Rendezvous is an abstraction for passing a Tensor
// from a producer to a consumer, where the consumer may safely
// request the Tensor before or after it has been produced.  A
// producer never blocks when using a Rendezvous.  A consumer has the
// choice of making a blocking call or providing a callback: in either
// case, the consumer receives the Tensor as soon as it is available.

最低0.47元/天解锁文章