Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add call stack info during compile time #19067

Merged
merged 4 commits into from Aug 8, 2019
Merged

Add call stack info during compile time #19067

merged 4 commits into from Aug 8, 2019

Conversation

chengduoZH
Copy link
Contributor

@chengduoZH chengduoZH commented Aug 8, 2019 •
edited

This PR's work includes:

  1. Add call stack info during compile time.
  2. Extract the common code of run time and compile time to operator_call_stack.

Before

Traceback (most recent call last):
  File "./train_gpu_paddle.py", line 471, in <module>
    train(n_token, cutoffs)
  File "./train_gpu_paddle.py", line 254, in train
    decr_ratio=FLAGS.decr_ratio)
  File "/ssd3/transformer-xl-paddlepaddle_tmp/optimization.py", line 151, in optimization
    _, param_grads = optimizer.minimize(loss)
  File "</usr/local/lib/python2.7/dist-packages/decorator.pyc:decorator-gen-20>", line 2, in minimize
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/wrapped_decorator.py", line 25, in __impl__
    return wrapped_func(*args, **kwargs)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/dygraph/base.py", line 86, in __impl__
    return func(*args, **kwargs)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/optimizer.py", line 594, in minimize
    no_grad_set=no_grad_set)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/optimizer.py", line 493, in backward
    no_grad_set, callbacks)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/backward.py", line 706, in append_backward
    _append_backward_vars_(root_block, fwd_op_num, grad_to_var, grad_info_map)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/backward.py", line 518, in _append_backward_vars_
    op_desc.infer_shape(block.desc)
paddle.fluid.core_avx.EnforceNotMet: Invoke operator elementwise_add_grad error.

Input(Out@GRAD) should not be null at [../paddle/fluid/operators/elementwise/elementwise_op.h:289]
PaddlePaddle Call Stacks:
0       0x7fb76db1b35ap void paddle::platform::EnforceNotMet::Init<char const*>(char const*, char const*, int) + 506
1       0x7fb76db1c065p paddle::platform::EnforceNotMet::EnforceNotMet(std::__exception_ptr::exception_ptr, char const*, int) + 165
2       0x7fb76f427ebap paddle::operators::ElementwiseOpExplicitGrad::InferShape(paddle::framework::InferShapeContext*) const + 1130
3       0x7fb76dcdf605p
4       0x7fb76dce3ed9p paddle::framework::OpDesc::InferShape(paddle::framework::BlockDesc const&) const + 985
5       0x7fb76dc00334p
6       0x7fb76db4eb14p
7             0x4c5326p PyEval_EvalFrameEx + 37958
8             0x4b9b66p PyEval_EvalCodeEx + 774
9             0x4c1f56p PyEval_EvalFrameEx + 24694
10            0x4b9b66p PyEval_EvalCodeEx + 774
11            0x4c17c6p PyEval_EvalFrameEx + 22758
12            0x4b9b66p PyEval_EvalCodeEx + 774
13            0x4c17c6p PyEval_EvalFrameEx + 22758
14            0x4b9b66p PyEval_EvalCodeEx + 774
15            0x4d57a3p
16            0x4a587ep PyObject_Call + 62
17            0x4be51ep PyEval_EvalFrameEx + 9790
18            0x4b9b66p PyEval_EvalCodeEx + 774
19            0x4d57a3p

Now:

Traceback (most recent call last):
  File "./train_gpu_paddle.py", line 471, in <module>
    train(n_token, cutoffs)
  File "./train_gpu_paddle.py", line 254, in train
    decr_ratio=FLAGS.decr_ratio)
  File "/ssd3/transformer-xl-paddlepaddle_tmp/optimization.py", line 151, in optimization
    _, param_grads = optimizer.minimize(loss)
  File "</usr/local/lib/python2.7/dist-packages/decorator.pyc:decorator-gen-20>", line 2, in minimize
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/wrapped_decorator.py", line 25, in __impl__
    return wrapped_func(*args, **kwargs)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/dygraph/base.py", line 86, in __impl__
    return func(*args, **kwargs)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/optimizer.py", line 594, in minimize
    no_grad_set=no_grad_set)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/optimizer.py", line 493, in backward
    no_grad_set, callbacks)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/backward.py", line 706, in append_backward
    _append_backward_vars_(root_block, fwd_op_num, grad_to_var, grad_info_map)
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/backward.py", line 518, in _append_backward_vars_
    op_desc.infer_shape(block.desc)
paddle.fluid.core_avx.EnforceNotMet: Invoke operator elementwise_add_grad error.

Python Callstacks:
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/framework.py", line 1776, in append_op
    attrs=kwargs.get("attrs", None))
  File "/ssd3/zcd_Paddle/build_fast/python/paddle/fluid/layers/math_op_patch.py", line 149, in __impl__
    attrs={'axis': axis})
  File "/ssd3/transformer-xl-paddlepaddle_tmp/model.py", line 371, in gather_nd
    index_join = raw_index * tensor.shape[1] + col_index
  File "/ssd3/transformer-xl-paddlepaddle_tmp/model.py", line 403, in _gather_logprob
    return gather_nd(logprob, idx)
  File "/ssd3/transformer-xl-paddlepaddle_tmp/model.py", line 494, in mask_adaptive_logsoftmax
    _gather_logprob(tail_logprob, cur_target))
  File "/ssd3/transformer-xl-paddlepaddle_tmp/model.py", line 558, in transformer
    proj_same_dim=proj_same_dim)
  File "/ssd3/transformer-xl-paddlepaddle_tmp/transformer_xl.py", line 200, in _build_model
    name='encoder'
  File "./train_gpu_paddle.py", line 84, in model_fn
    loss, new_mems = transofrmer_xl._build_model()
  File "./train_gpu_paddle.py", line 99, in single_core_graph
    is_training=is_training)
  File "./train_gpu_paddle.py", line 128, in create_model
    mems=mems)
  File "./train_gpu_paddle.py", line 238, in train
    batch_size=bsz_per_trainner, n_token=n_token, cutoffs=cutoffs)
  File "./train_gpu_paddle.py", line 471, in <module>
    train(n_token, cutoffs)
C++ Callstacks:
Input(Out@GRAD) should not be null at [../paddle/fluid/operators/elementwise/elementwise_op.h:289]
PaddlePaddle Call Stacks:
0       0x7fb76db1b35ap void paddle::platform::EnforceNotMet::Init<char const*>(char const*, char const*, int) + 506
1       0x7fb76db1c065p paddle::platform::EnforceNotMet::EnforceNotMet(std::__exception_ptr::exception_ptr, char const*, int) + 165
2       0x7fb76f427ebap paddle::operators::ElementwiseOpExplicitGrad::InferShape(paddle::framework::InferShapeContext*) const + 1130
3       0x7fb76dcdf605p
4       0x7fb76dce3ed9p paddle::framework::OpDesc::InferShape(paddle::framework::BlockDesc const&) const + 985
5       0x7fb76dc00334p
6       0x7fb76db4eb14p
7             0x4c5326p PyEval_EvalFrameEx + 37958
8             0x4b9b66p PyEval_EvalCodeEx + 774
9             0x4c1f56p PyEval_EvalFrameEx + 24694
10            0x4b9b66p PyEval_EvalCodeEx + 774
11            0x4c17c6p PyEval_EvalFrameEx + 22758
12            0x4b9b66p PyEval_EvalCodeEx + 774

@chengduoZH chengduoZH force-pushed the add_error_info_during_compile branch from 11f2c87 to f5590b1 Compare August 8, 2019 03:36
@chengduoZH chengduoZH requested a review from Xreki August 8, 2019 03:41
@chengduoZH chengduoZH force-pushed the add_error_info_during_compile branch from 10d56aa to 9a8997f Compare August 8, 2019 06:05
Add unit test
efab284 
test=develop
chenwhql
chenwhql reviewed Aug 8, 2019
View reviewed changes
Copy link
Contributor

@chenwhql chenwhql left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

just my thinking, maybe the Traceback info and C++ callstacks info is useless for most users, because they just use our platform, they will not modify our source code, they just need to know where their own mistakes are. The error message is too long now.

chenwhql
chenwhql reviewed Aug 8, 2019
View reviewed changes
python/paddle/fluid/tests/unittests/test_runtime_and_compiletime_exception.py Outdated


class TestCompileTimeException(OpTest):
def test_run_time_exception(self):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

maybe test_compile_time_exception is better?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done

follow comment
142d30c 
test=develop
@chengduoZH
Copy link
Contributor Author

just my thinking, maybe the Traceback info and C++ callstacks info is useless for most users, because they just use our platform, they will not modify our source code, they just need to know where their own mistakes are. The error message is too long now.

C++ Callstack info is really long, but some error info of C++ Callstack is useful for users and developers.

chenwhql
chenwhql approved these changes Aug 8, 2019
View reviewed changes
Copy link
Contributor

@chenwhql chenwhql left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@chengduoZH chengduoZH merged commit 21440b4 into PaddlePaddle:develop Aug 8, 2019
@chengduoZH chengduoZH mentioned this pull request Aug 20, 2019
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@chenwhql chenwhql chenwhql approved these changes

@Xreki Xreki Awaiting requested review from Xreki

Assignees
No one assigned
Labels
Error Message
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
3 participants
@chengduoZH @chenwhql @luotao1