Revert "Revert D28833086: beef up at::_ops API" #60214

bdhirsh · 2021-06-17T22:21:21Z

Reland of #59115, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871)

This is identical to the original PR except for one change in tools/codegen/gen.py: static constexpr -> static CONSTEXPR_CONST_EXCEPT_WIN_CUDA (had to make a new macro that was a slight adjustment to an existing one)

This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675

This reverts commit 6d0fb85.

Stack from ghstack:

Revert "Revert D28833086: beef up at::_ops API" #60214 Revert "Revert D28833086: beef up at::_ops API"

Differential Revision: D29213932

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. [ghstack-poisoned]

facebook-github-bot · 2021-06-17T22:21:28Z

💊 CI failures summary and remediations

As of commit 902b609 (more details on the Dr. CI page and at hud.pytorch.org/pr/60214):

3/4 failures introduced in this PR
1/4 broken upstream at merge base acf791e on Jun 17 from 1:09pm to 7:18pm

🕵️ 3 new failures recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

pytorch_xla_linux_bionic_py3_6_clang9_test (1/3)

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

Jun 24 18:24:05 AssertionError: "weight tensor ...nsion is 4, the 1th output dimension is 3. vs. OK)

Jun 24 18:24:05 *** End stack trace ***
Jun 24 18:24:05 
Jun 24 18:24:05 
Jun 24 18:24:05 During handling of the above exception, another exception occurred:
Jun 24 18:24:05 
Jun 24 18:24:05 Traceback (most recent call last):
Jun 24 18:24:05   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_device_type.py", line 397, in instantiated_test
Jun 24 18:24:05     result = test_fn(self, *args)
Jun 24 18:24:05   File "/var/lib/jenkins/workspace/xla/test/../../test/test_nn.py", line 16007, in test_nll_loss_invalid_weights
Jun 24 18:24:05     F.nll_loss(x, t, weight=weight)
Jun 24 18:24:05 AssertionError: "weight tensor should be defined either for all 3 classes or no classes" does not match "/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/tensorflow/compiler/xla/xla_client/debug_macros.h:27 : Check failed: status.status() == ::tensorflow::Status::OK() (Invalid argument: Input dimension should be either 1 or equal to the output dimension it is broadcasting into; the 0th operand dimension is 4, the 1th output dimension is 3. vs. OK)
Jun 24 18:24:05 *** Begin stack trace ***
Jun 24 18:24:05 	tensorflow::CurrentStackTrace[abi:cxx11]()
Jun 24 18:24:05 	xla::Shape const* ConsumeValue<xla::Shape const*>(tensorflow::StatusOr<xla::Shape const*>&&)
Jun 24 18:24:05 	torch_xla::XlaHelpers::ShapeOfXlaOp(xla::XlaOp)
Jun 24 18:24:05 	torch_xla::ir::ops::InferOutputShape(absl::lts_20210324::Span<xla::Shape const>, std::function<xla::XlaOp (absl::lts_20210324::Span<xla::XlaOp const>)> const&)
Jun 24 18:24:05 	
Jun 24 18:24:05 	torch_xla::ir::Node::GetOpShape(std::function<xla::Shape ()> const&) const
Jun 24 18:24:05 	torch_xla::ir::Node::Node(torch_xla::ir::OpKind, absl::lts_20210324::Span<torch_xla::ir::Value const>, std::function<xla::Shape ()> const&, unsigned long, absl::lts_20210324::uint128)
Jun 24 18:24:05 	torch_xla::ir::ops::NllLoss::NllLoss(torch_xla::ir::Value const&, torch_xla::ir::Value const&, absl::lts_20210324::optional<torch_xla::ir::Value> const&, torch_xla::ReductionMode, int)
Jun 24 18:24:05 	torch_xla::XLATensor::nll_loss(torch_xla::XLATensor const&, torch_xla::XLATensor const&, torch_xla::XLATensor const&, long, int)

binary_windows_wheel_3_7_cu102_build (2/3)

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

ERROR: Command errored out with exit status 1: ...pmgkyp1ij' Check the logs for full command output.

      config_settings, requirements=['wheel'])
    File "C:\Users\circleci\AppData\Local\Temp\pip-build-env-rru106ve\overlay\Lib\site-packages\setuptools\build_meta.py", line 135, in _get_build_requires
      self.run_setup()
    File "C:\Users\circleci\AppData\Local\Temp\pip-build-env-rru106ve\overlay\Lib\site-packages\setuptools\build_meta.py", line 150, in run_setup
      exec(compile(code, __file__, 'exec'), locals())
    File "setup.py", line 219, in <module>
      from tools.build_pytorch_libs import build_caffe2
  ModuleNotFoundError: No module named 'tools'
  Getting requirements to build wheel: finished with status 'error'
WARNING: Discarding file:///C:/w/b/windows/pytorch. Command errored out with exit status 1: 'C:\w\b\windows\conda\envs\py37\python.exe' 'C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_vendor\pep517\in_process\_in_process.py' get_requires_for_build_wheel 'C:\Users\circleci\AppData\Local\Temp\tmpmgkyp1ij' Check the logs for full command output.
ERROR: Command errored out with exit status 1: 'C:\w\b\windows\conda\envs\py37\python.exe' 'C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_vendor\pep517\in_process\_in_process.py' get_requires_for_build_wheel 'C:\Users\circleci\AppData\Local\Temp\tmpmgkyp1ij' Check the logs for full command output.
Exception information:
Traceback (most recent call last):
  File "C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_internal\cli\base_command.py", line 180, in _main
    status = self.run(options, args)
  File "C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_internal\cli\req_command.py", line 204, in wrapper
    return func(self, options, args)
  File "C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_internal\commands\wheel.py", line 143, in run
    reqs, check_supported_wheels=True
  File "C:\w\b\windows\conda\envs\py37\lib\site-packages\pip\_internal\resolution\resolvelib\resolver.py", line 104, in resolve
    req, requested_extras=()

pytorch_libtorch_linux_xenial_cuda10_2_cudnn7_py3_gcc7_build (3/3)

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

Jun 24 15:04:47 rm: cannot remove '/var/lib/jenkins/sccache_error.log': No such file or directory

Jun 24 15:04:47 ++++ extract_trap_cmd
Jun 24 15:04:47 ++++ printf '%s\n' ''
Jun 24 15:04:47 +++ printf '%s\n' cleanup
Jun 24 15:04:47 ++ trap -- '
Jun 24 15:04:47 cleanup' EXIT
Jun 24 15:04:47 ++ [[ pytorch-libtorch-linux-xenial-cuda10.2-cudnn7-py3-gcc7-build != *pytorch-win-* ]]
Jun 24 15:04:47 ++ which sccache
Jun 24 15:04:47 ++ sccache --stop-server
Jun 24 15:04:47 ++ true
Jun 24 15:04:47 ++ rm /var/lib/jenkins/sccache_error.log
Jun 24 15:04:47 rm: cannot remove '/var/lib/jenkins/sccache_error.log': No such file or directory
Jun 24 15:04:47 ++ true
Jun 24 15:04:47 ++ [[ -n '' ]]
Jun 24 15:04:47 ++ [[ pytorch-libtorch-linux-xenial-cuda10.2-cudnn7-py3-gcc7-build == *rocm* ]]
Jun 24 15:04:47 ++ SCCACHE_ERROR_LOG=/var/lib/jenkins/sccache_error.log
Jun 24 15:04:47 ++ SCCACHE_IDLE_TIMEOUT=1200
Jun 24 15:04:47 ++ RUST_LOG=sccache::server=error
Jun 24 15:04:47 ++ sccache --start-server
Jun 24 15:04:47 sccache: Starting the server...
Jun 24 15:04:47 ++ sccache --zero-stats
Jun 24 15:04:47 Compile requests                      0

2 jobs timed out:

pytorch_libtorch_linux_xenial_cuda11_1_cudnn8_py3_gcc7_build
pytorch_libtorch_linux_xenial_cuda10_2_cudnn7_py3_gcc7_build

🚧 1 fixed upstream failure:

These were probably caused by upstream breakages that were already fixed.

Please rebase on the viable/strict branch (expand for instructions)

If your commit is older than viable/strict, run these commands:

git fetch https://github.com/pytorch/pytorch viable/strict
git rebase FETCH_HEAD

pytorch_libtorch_linux_xenial_cuda11_1_cudnn8_py3_gcc7_build on Jun 17 from 1:09pm to 7:18pm (e2129d1 - e2129d1)
- 🔁 rerun

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 to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

bdhirsh · 2021-06-17T22:24:13Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_EXCEPT_WIN_CUDA` This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: f6b3d0f Pull Request resolved: #60214

bdhirsh · 2021-06-17T23:52:36Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

bdhirsh · 2021-06-18T19:21:19Z

update 1: I had to make a slightly new macro to handle constexpr const properly: CONSTEXPR_CONST_EXCEPT_WIN_CUDA

update 2: looks like in the brief period of time that my original PR landed, some small binary size regressions were detected internally. Gotta figure out where that was coming from before I re-land this. Although inspecting the OSS libtorch_cpu.so showed that we were correctly adding two new symbols for each new op, at::_ops::add_Tensor::call and at::_ops::add_Tensor::redispatch and removing/inlining the corresponding previous symbols, at::add and at::redispatch::add. Maybe internal testing will help more.

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: ac6ce82 Pull Request resolved: #60214

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 514fa0b Pull Request resolved: #60214

bdhirsh · 2021-06-22T13:59:31Z

Okay, issue (1) is harder to solve than I thought.

My proposed fix (it's in the PR)

For everything except Windows NVCC builds we do this:

struct add_Tensor {
    static constexpr const char* name = "add";
};

And for Windows NVCC builds we do this:

struct add_Tensor {
    // const instead of constexpr
    static const char* name;
};
// defined outside of the struct definition
const char* add_Tensor name = "add";

I added a note with more details in Macros.h

The problem

I'd like to write code like the following:

struct add_Tensor {
    static constexpr const char* name = "add";
};

But Windows NVCC apparently has a bug (in VS2017, fixed in VS2019) where it can't deal with static constexpr. A workaround issue recommends using static const instead (https://developercommunity.visualstudio.com/t/static-constexpr-gives-member-may-not-be-initializ/444344).

The next reasonable alternative to try is this (we actually have some code that does this with an existing CONSTEXPR_EXCEPT_WIN_CUDA macro):

struct add_Tensor {
    static const char* name = "add";
};

The above fails thanks to the C++ standard (see https://developercommunity.visualstudio.com/t/static-constexpr-gives-member-may-not-be-initializ/444344).

I tried out those two options and a few others in Compiler Explorer - see the examples with descriptions in https://godbolt.org/z/Tn73xdYGz.

It would be great if we can suppress the original Windows NVCC compiler error, but I haven't figured out a way to do that. The alternative that I added at the top works fine, but accesses to at::_ops::add_Tensor::name will be a little slower for windows-NVCC, so we just need to make sure it's not used anywhere that's perf critical.

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

bdhirsh · 2021-06-22T15:47:15Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 2863525 Pull Request resolved: #60214

ezyang · 2021-06-22T17:09:29Z

This is fine, string access is guarded behind executed only once static calls anyway

Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

bdhirsh · 2021-06-22T20:45:42Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Reland of #59115, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 6c979dc Pull Request resolved: #60214

bdhirsh · 2021-06-22T22:18:07Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Reland of #59115, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 9f9363a Pull Request resolved: #60214

bdhirsh · 2021-06-22T22:43:20Z

@bdhirsh has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

bdhirsh · 2021-06-23T21:37:44Z

Still waiting for some internal binary size tests to finish running, but after talking to @ezyang I moved some incriminating functions that I'd previously moved from Functions.cpp and TensorMethods.cpp into Functions.h and TensorBody.h back into their original files, and so far I haven't seen any failed tests.

There were a few functions that I'd moved into the header files before that I'm keeping in for now, like Tensor::cpu() and Tensor::options(), because they seem small and useful enough that we might get some perf wins without sacrificing too much in terms of binary size. If more tests fail then I'll try moving those back to cpp files too.

Reland of #59115, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

Pull Request resolved: #60214 Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_EXCEPT_WIN_CUDA` This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D29213932/)! ghstack-source-id: 132130129

Reland of #59115, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_CONST_EXCEPT_WIN_CUDA` (had to make a new macro that was a slight adjustment to an existing one) This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932) [ghstack-poisoned]

Pull Request resolved: #60214 Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_EXCEPT_WIN_CUDA` This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Differential Revision: [D29213932](https://our.internmc.facebook.com/intern/diff/D29213932/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D29213932/)! ghstack-source-id: 132300710

facebook-github-bot · 2021-06-25T01:10:34Z

@bdhirsh merged this pull request in 7bc8645.

Summary: Pull Request resolved: pytorch#60214 Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_EXCEPT_WIN_CUDA` This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: pytorch#40675 This reverts commit 6d0fb85. Test Plan: Imported from OSS Reviewed By: ezyang Differential Revision: D29213932 Pulled By: bdhirsh fbshipit-source-id: b90c7c10e5a51f8d6173ddca673b418e5774c248

Summary: Pull Request resolved: #60214 Relanding this PR, but with a fix for windows cuda builds (example failure in master here: https://github.com/pytorch/pytorch/runs/2852662871) This is identical to the original PR except for one change in `tools/codegen/gen.py`: `static constexpr` -> `static CONSTEXPR_EXCEPT_WIN_CUDA` This actually took a while to figure out, until I tracked down a previous pytorch PR that encountered a similar issue: #40675 This reverts commit 6d0fb85. Test Plan: Imported from OSS Reviewed By: ezyang Differential Revision: D29213932 Pulled By: bdhirsh fbshipit-source-id: b90c7c10e5a51f8d6173ddca673b418e5774c248

Revert "Revert D28833086: beef up at::_ops API"

c8d5664

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. [ghstack-poisoned]

facebook-github-bot added the cla signed label Jun 17, 2021

bdhirsh requested a review from ezyang June 17, 2021 22:23

bdhirsh requested a review from bhosmer June 17, 2021 22:25

bdhirsh added ci/all and removed ci/all labels Jun 17, 2021

bdhirsh added a commit that referenced this pull request Jun 17, 2021

Revert "Revert D28833086: beef up at::_ops API"

ea8691c

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: f6b3d0f Pull Request resolved: #60214

ezyang approved these changes Jun 18, 2021

View reviewed changes

bdhirsh added a commit that referenced this pull request Jun 18, 2021

Revert "Revert D28833086: beef up at::_ops API"

523ba1c

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: ac6ce82 Pull Request resolved: #60214

bhosmer approved these changes Jun 21, 2021

View reviewed changes

bdhirsh added a commit that referenced this pull request Jun 22, 2021

Revert "Revert D28833086: beef up at::_ops API"

7f937c7

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 514fa0b Pull Request resolved: #60214

bdhirsh added a commit that referenced this pull request Jun 22, 2021

Revert "Revert D28833086: beef up at::_ops API"

4ec70c1

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 2863525 Pull Request resolved: #60214

bdhirsh added a commit that referenced this pull request Jun 22, 2021

Revert "Revert D28833086: beef up at::_ops API"

bf75579

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 6c979dc Pull Request resolved: #60214

bdhirsh added a commit that referenced this pull request Jun 22, 2021

Revert "Revert D28833086: beef up at::_ops API"

40d3b75

relanding this PR, but with a fix for windows cuda builds This reverts commit 6d0fb85. ghstack-source-id: 9f9363a Pull Request resolved: #60214

facebook-github-bot closed this in 7bc8645 Jun 25, 2021

facebook-github-bot added the Merged label Jun 25, 2021

facebook-github-bot deleted the gh/bdhirsh/126/head branch June 28, 2021 14:17

Revert "Revert D28833086: beef up at::_ops API" #60214

Revert "Revert D28833086: beef up at::_ops API" #60214

Uh oh!

Conversation

bdhirsh commented Jun 17, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

facebook-github-bot commented Jun 17, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

💊 CI failures summary and remediations

🕵️ 3 new failures recognized by patterns

pytorch_xla_linux_bionic_py3_6_clang9_test (1/3)

binary_windows_wheel_3_7_cu102_build (2/3)

pytorch_libtorch_linux_xenial_cuda10_2_cudnn7_py3_gcc7_build (3/3)

🚧 1 fixed upstream failure:

Uh oh!

bdhirsh commented Jun 17, 2021

Uh oh!

bdhirsh commented Jun 17, 2021

Uh oh!

bdhirsh commented Jun 18, 2021

Uh oh!

bdhirsh commented Jun 22, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bdhirsh commented Jun 22, 2021

Uh oh!

ezyang commented Jun 22, 2021

Uh oh!

bdhirsh commented Jun 22, 2021

Uh oh!

bdhirsh commented Jun 22, 2021

Uh oh!

bdhirsh commented Jun 22, 2021

Uh oh!

bdhirsh commented Jun 23, 2021

Uh oh!

facebook-github-bot commented Jun 25, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

bdhirsh commented Jun 17, 2021 •

edited

Loading

facebook-github-bot commented Jun 17, 2021 •

edited

Loading

bdhirsh commented Jun 22, 2021 •

edited

Loading