max_memory_reserved¶

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

Return the peak size of GPU memory that is held by the allocator of the given device.

Parameters

device (paddle.CUDAPlace or int or str, optional) – The device, the id of the device or the string name of device like ‘gpu:x’. If device is None, the device is the current device. Default: None.

Returns

The peak size of GPU memory that is held by the allocator of the given device, in bytes.

Return type

int

Examples

# required: gpu
import paddle

max_memory_reserved_size = paddle.device.cuda.max_memory_reserved(paddle.CUDAPlace(0))
max_memory_reserved_size = paddle.device.cuda.max_memory_reserved(0)
max_memory_reserved_size = paddle.device.cuda.max_memory_reserved("gpu:0")