torch.xpu.memory_allocated¶
- torch.xpu.memory_allocated(device=None)[source]¶
Return the current GPU memory occupied by tensors in bytes for a given device.
- Parameters
device (torch.device or int or str, optional) – selected device. Returns statistic for the current device, given by
current_device()
, ifdevice
isNone
(default).- Return type
Note
This is likely less than the amount shown in xpu-smi since some unused memory can be held by the caching allocator and some context needs to be created on GPU.