Add timeout and loc for blocking calls

[jackalcooper]

No description provided.

[jackalcooper]

效果：

E0824 10:59:54.763103 4013850 tensor.cpp:96] [rank=1] blocking detected. Python stack:
  File "graph/train_consistent.py", line 370, in <module>
    main(args)

  File "graph/train_consistent.py", line 310, in main
    l = loss.numpy()

  File "/home/caishenghang/oneflow_cuda/python/oneflow/framework/tensor.py", line 50, in _tensor_numpy
    with BlockingContext() as ctx:

  File "/home/caishenghang/oneflow_cuda/python/oneflow/support/blocking.py", line 9, in __init__
    stack_info = "\n".join(traceback.format_stack(limit=5))
E0824 10:59:57.762894 4013849 transport_util.cpp:34] [rank=0] blocking detected. Python stack:
  File "graph/train_consistent.py", line 370, in <module>
    main(args)

  File "graph/train_consistent.py", line 348, in main
    flow.save(

  File "/home/caishenghang/oneflow_cuda/python/oneflow/framework/check_point_v2.py", line 166, in save
    var_dict[name] = var.to_consistent(sbp=flow.sbp.broadcast).to_local()

  File "/home/caishenghang/oneflow_cuda/python/oneflow/nn/modules/consistent_cast.py", line 104, in to_consistent_op
    with BlockingContext() as ctx:

  File "/home/caishenghang/oneflow_cuda/python/oneflow/support/blocking.py", line 9, in __init__
    stack_info = "\n".join(traceback.format_stack(limit=5))

[lixinqi]

BlockingInfoContext

[lixinqi]

这里不应该被注释。

[lixinqi]

这里的heartbeat_interval_seconds是不是应该是seconds

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

PyTorch resnet50 time: 139.8ms (= 6989.5ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 127.9ms (= 6395.8ms / 50, input_shape=[16, 3, 224, 224], backward is enabled)
Relative speed: 1.09 (= 139.8ms / 127.9ms)

PyTorch resnet50 time: 84.4ms (= 4218.3ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 74.4ms (= 3720.6ms / 50, input_shape=[8, 3, 224, 224], backward is enabled)
Relative speed: 1.13 (= 84.4ms / 74.4ms)

PyTorch resnet50 time: 57.8ms (= 2888.6ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 50.2ms (= 2509.5ms / 50, input_shape=[4, 3, 224, 224], backward is enabled)
Relative speed: 1.15 (= 57.8ms / 50.2ms)

PyTorch resnet50 time: 45.4ms (= 2269.3ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 43.4ms (= 2169.6ms / 50, input_shape=[2, 3, 224, 224], backward is enabled)
Relative speed: 1.05 (= 45.4ms / 43.4ms)

PyTorch resnet50 time: 42.7ms (= 2136.8ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
OneFlow resnet50 time: 40.4ms (= 2022.1ms / 50, input_shape=[1, 3, 224, 224], backward is enabled)
Relative speed: 1.06 (= 42.7ms / 40.4ms)