caching_allocator_alloc(size, device=None, stream=None)[source]¶
Performs a memory allocation using the CUDA memory allocator.
Memory is allocated for a given device and a stream, this function is intended to be used for interoperability with other frameworks. Allocated memory is released through
size (int) – number of bytes to be allocated.
device (torch.device or int, optional) – selected device. If it is
Nonethe default CUDA device is used.
stream (torch.cuda.Stream or int, optional) – selected stream. If is
Nonethen the default stream for the selected device is used.
See Memory management for more details about GPU memory management.