MCPcopy Create free account
hub / github.com/pytorch/pytorch / stream

Method stream

torch/_dynamo/device_interface.py:81–82  ·  view source on GitHub ↗
(stream: torch.Stream)

Source from the content-addressed store, hash-verified

79
80 @staticmethod
81 def stream(stream: torch.Stream):
82 raise NotImplementedError()
83
84 @staticmethod
85 def current_stream():

Callers 15

to_mapFunction · 0.80
_refillMethod · 0.80
_refillMethod · 0.80
_unshardFunction · 0.80
_post_backward_hookFunction · 0.80
_reduce_gradFunction · 0.80
use_streamFunction · 0.80
apply_optim_in_backward_hookFunction · 0.80
_reducer_allreduce_and_upcast_hookFunction · 0.80
_use_cuda_memory_pool_managerFunction · 0.80
cudagraphify_implFunction · 0.80
_workerFunction · 0.80

Calls

no outgoing calls

Tested by 15

test_DistributedDataParallel_non_default_streamMethod · 0.64
async_cuda_sleep_and_set_to_oneFunction · 0.64
forwardMethod · 0.64
_test_cuda_future_extractionMethod · 0.64
_slow_add_on_user_streamMethod · 0.64
test_cuda_future_callback_changes_devicesMethod · 0.64
test_cuda_future_value_on_bad_deviceMethod · 0.64
test_async_execution_with_cuda_futureMethod · 0.64
test_async_execution_nested_with_cuda_futureMethod · 0.64
test_fused_attention_vs_math_ref_grads_cudagraphMethod · 0.64
cuda_multiply_twoFunction · 0.64
test_eventMethod · 0.64