synchronize

paddle.device.cuda. synchronize ( device=None ) [source]

Wait for the compute on the given CUDA device to finish.

Parameters

device (paddle.CUDAPlace()|int, optional) – The device or the ID of the device. If device is None, the device is the current device. Default: None.

Examples

# required: gpu
import paddle

paddle.device.cuda.synchronize()
paddle.device.cuda.synchronize(0)
paddle.device.cuda.synchronize(paddle.CUDAPlace(0))