Don't double dispatch on CUDA logical_not. by gchanan · Pull Request #38275 · pytorch/pytorch

gchanan · 2020-05-11T21:56:18Z

Stack from ghstack:

Don't double dispatch on CUDA logical_not. #38275 Don't double dispatch on CUDA logical_not.
Move Cuda Abs kernel to its own file. #38274 Move Cuda Abs kernel to its own file.

This op doesn't need to be fast and the double dispatch is time consuming to compile.

Differential Revision: D21511832

This op doesn't need to be fast and the double dispatch is time consuming to compile. [ghstack-poisoned]

This op doesn't need to be fast and the double dispatch is time consuming to compile. ghstack-source-id: 90bd012 Pull Request resolved: #38275

gchanan · 2020-05-11T22:00:20Z

I think(?) this has the same behavior and the cases I tried have the same behavior, but it would be good to confirm.

dr-ci · 2020-05-11T22:08:32Z

💊 CI failures summary and remediations

As of commit 3c818a0 (more details on the Dr. CI page):

1/3 failures possibly* introduced in this PR
- 1/1 non-CircleCI failure(s)
2/3 tentatively recognized as flaky ❄️
- Click here to rerun these jobs

❄️ 2 failures tentatively classified as flaky

but reruns have not yet been triggered to confirm:

pytorch_windows_vs2019_py36_cuda10.1_test2 (1/2)

Step: "Test" (full log | diagnosis details | 🔁 rerun) ❄️

AssertionError: 1 not less than or equal to 0 :

  File "C:\Users\circleci\project\build\win_tmp\build\torch\testing\_internal\common_device_type.py", line 234, in instantiated_test 
    result = test(self, device_arg, dtype) 
  File "test_torch.py", line 6458, in test_logical_not_out 
    self.assertEqual(out_np, out.to('cpu')) 
  File "C:\Users\circleci\project\build\win_tmp\build\torch\testing\_internal\common_utils.py", line 1010, in assertEqual 
    allow_inf=allow_inf, exact_dtype=exact_dtype) 
  File "C:\Users\circleci\project\build\win_tmp\build\torch\testing\_internal\common_utils.py", line 913, in assertEqual 
    allow_inf=allow_inf, exact_dtype=exact_dtype) 
  File "C:\Users\circleci\project\build\win_tmp\build\torch\testing\_internal\common_utils.py", line 1020, in assertEqual 
    super().assertLessEqual(abs(x - y), atol, message) 
AssertionError: 1 not less than or equal to 0 :  
 
---------------------------------------------------------------------- 
Ran 5484 tests in 367.612s 
 
FAILED (failures=45, skipped=318) 
 
Generating XML reports... 
Generated XML report: test-reports\python-unittest\TEST-TestDevicePrecisionCUDA-20200514213619.xml 
Generated XML report: test-reports\python-unittest\TEST-TestTensorDeviceOpsCPU-20200514213619.xml 
Generated XML report: test-reports\python-unittest\TEST-TestTensorDeviceOpsCUDA-20200514213619.xml

pytorch_linux_xenial_cuda10_2_cudnn7_py3_gcc7_test (2/2)

Step: "Run tests" (full log | diagnosis details | 🔁 rerun) ❄️

May 14 22:03:15 AssertionError: 1 not less than or equal to 0 :

May 14 22:03:15   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_device_type.py", line 234, in instantiated_test 
May 14 22:03:15     result = test(self, device_arg, dtype) 
May 14 22:03:15   File "test_torch.py", line 6458, in test_logical_not_out 
May 14 22:03:15     self.assertEqual(out_np, out.to('cpu')) 
May 14 22:03:15   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_utils.py", line 1010, in assertEqual 
May 14 22:03:15     allow_inf=allow_inf, exact_dtype=exact_dtype) 
May 14 22:03:15   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_utils.py", line 913, in assertEqual 
May 14 22:03:15     allow_inf=allow_inf, exact_dtype=exact_dtype) 
May 14 22:03:15   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_utils.py", line 1020, in assertEqual 
May 14 22:03:15     super().assertLessEqual(abs(x - y), atol, message) 
May 14 22:03:15 AssertionError: 1 not less than or equal to 0 :  
May 14 22:03:15  
May 14 22:03:16 ---------------------------------------------------------------------- 
May 14 22:03:16 Ran 5484 tests in 429.714s 
May 14 22:03:16  
May 14 22:03:16 FAILED (failures=45, skipped=367) 
May 14 22:03:16  
May 14 22:03:16 Generating XML reports... 
May 14 22:03:16 Generated XML report: test-reports/dist-gloo/TEST-TestDevicePrecisionCUDA-20200514215604.xml 
May 14 22:03:16 Generated XML report: test-reports/dist-gloo/TEST-TestTensorDeviceOpsCPU-20200514215604.xml 
May 14 22:03:16 Generated XML report: test-reports/dist-gloo/TEST-TestTensorDeviceOpsCUDA-20200514215604.xml

ci.pytorch.org: 1 failed

Failed: pr/py3.6-clang7-rocmdeb-ubuntu16.04

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker.

See how this bot performed.

This comment has been revised 57 times.

mruberry · 2020-05-11T22:09:32Z

The logical_not test

pytorch/test/test_torch.py

Line 6399 in a553935

def test_logical_not(self, device):

is pretty milquetoast and doesn't use our latest @dtypes gadgetry or compare against NumPy. Want to take this opportunity to update it?

mruberry · 2020-05-11T22:10:20Z

The logical_not test

pytorch/test/test_torch.py

Line 6399 in a553935

def test_logical_not(self, device):

is pretty milquetoast and doesn't use our latest @dtypes gadgetry or compare against NumPy. Want to take this opportunity to update it?

mruberry

The logical_not test

pytorch/test/test_torch.py

Line 6399 in a553935

def test_logical_not(self, device):

is pretty milquetoast and doesn't use our latest @dtypes gadgetry or compare against NumPy. Want to take this opportunity to update it?