Port `prod` kernel to structured kernels. #62024

ysiraichi · 2021-07-22T10:00:35Z

Stack from ghstack:

Port norm kernel to structured kernels. #62711 Port norm kernel to structured kernels.
Port prod kernel to structured kernels. #62024 Port prod kernel to structured kernels.
Port mean kernel to structured kernels. #61643 Port mean kernel to structured kernels.

Tracking issue: #55070

In this PR, I also broke down the meta functions of other reduction kernels (e.g. all,
argmax, sum) into the composition of common patterns.

Differential Revision: D29847122

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. [ghstack-poisoned]

facebook-github-bot · 2021-07-22T10:00:41Z

🔗 Helpful links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/62024
📄 Preview docs built from this PR

💊 CI failures summary and remediations

As of commit 4dc4316 (more details on the Dr. CI page):

3/4 failures introduced in this PR
1/4 tentatively recognized as flaky ❄️
- Click here to rerun these jobs

🕵️ 2 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_build (1/2)

Step: "(Optional) Merge target branch" (full log | diagnosis details | 🔁 rerun)

Automatic merge failed; fix conflicts and then commit the result.

CONFLICT (add/add): Merge conflict in torch/csrc/autograd/profiler_utils.cpp
Auto-merging torch/csrc/autograd/profiler_utils.cpp
CONFLICT (add/add): Merge conflict in test/cpp/jit/test_misc.cpp
Auto-merging test/cpp/jit/test_misc.cpp
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/quantized/cpu/qreduction.cpp
Auto-merging aten/src/ATen/native/quantized/cpu/qreduction.cpp
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/native_functions.yaml
Auto-merging aten/src/ATen/native/native_functions.yaml
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/ReduceOps.cpp
Auto-merging aten/src/ATen/native/ReduceOps.cpp
Automatic merge failed; fix conflicts and then commit the result.


Exited with code exit status 1

pytorch_linux_xenial_py3_6_gcc5_4_build (2/2)

Step: "(Optional) Merge target branch" (full log | diagnosis details | 🔁 rerun)

Automatic merge failed; fix conflicts and then commit the result.

HEAD is now at 4dc4316ce6 Update on "Port `prod` kernel to structured kernels."
+ git reset --hard 4dc4316ce64530ce7d7813a50970cf4af6dd4011
HEAD is now at 4dc4316ce6 Update on "Port `prod` kernel to structured kernels."
+ git merge --allow-unrelated-histories --no-edit --no-ff fb4ba9e6640a7f94658da45cc4c48e542849ff5f
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/quantized/cpu/qreduction.cpp
Auto-merging aten/src/ATen/native/quantized/cpu/qreduction.cpp
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/native_functions.yaml
Auto-merging aten/src/ATen/native/native_functions.yaml
CONFLICT (add/add): Merge conflict in aten/src/ATen/native/ReduceOps.cpp
Auto-merging aten/src/ATen/native/ReduceOps.cpp
Automatic merge failed; fix conflicts and then commit the result.


Exited with code exit status 1

1 failure not recognized by patterns:

Job	Step	Action
^{pytorch_linux_bionic_cuda10_2_cudnn7_py3_9_gcc7_test2}	^{Set Up CI Environment After attach_workspace}	🔁 rerun

❄️ 1 failure tentatively classified as flaky

but reruns have not yet been triggered to confirm:

linux-bionic-py3.8-gcc9-coverage / test (default, 2, 2, linux.2xlarge) (1/1)

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

2021-08-11T20:43:51.0353067Z RuntimeError: Proc...ated or timed out after 100.07439947128296 seconds

2021-08-11T20:43:51.0343824Z ======================================================================
2021-08-11T20:43:51.0344784Z ERROR [100.094s]: test_return_local_rrefs (__main__.TensorPipeRpcTestWithSpawn)
2021-08-11T20:43:51.0346563Z ----------------------------------------------------------------------
2021-08-11T20:43:51.0347187Z Traceback (most recent call last):
2021-08-11T20:43:51.0348445Z   File "/opt/conda/lib/python3.8/site-packages/torch/testing/_internal/common_distributed.py", line 418, in wrapper
2021-08-11T20:43:51.0349362Z     self._join_processes(fn)
2021-08-11T20:43:51.0350224Z   File "/opt/conda/lib/python3.8/site-packages/torch/testing/_internal/common_distributed.py", line 645, in _join_processes
2021-08-11T20:43:51.0350946Z     self._check_return_codes(elapsed_time)
2021-08-11T20:43:51.0351849Z   File "/opt/conda/lib/python3.8/site-packages/torch/testing/_internal/common_distributed.py", line 695, in _check_return_codes
2021-08-11T20:43:51.0352528Z     raise RuntimeError(
2021-08-11T20:43:51.0353067Z RuntimeError: Process 0 terminated or timed out after 100.07439947128296 seconds
2021-08-11T20:43:51.0353450Z 
2021-08-11T20:43:51.0353943Z ----------------------------------------------------------------------
2021-08-11T20:43:51.0354376Z Ran 369 tests in 757.239s
2021-08-11T20:43:51.0381459Z 
2021-08-11T20:43:51.0382078Z FAILED (errors=1, skipped=9)
2021-08-11T20:43:51.0382402Z 
2021-08-11T20:43:51.0383009Z Generating XML reports...
2021-08-11T20:43:51.0384906Z Generated XML report: test-reports/python-unittest/distributed.rpc.test_tensorpipe_agent/TEST-TensorPipeDdpComparisonTestWithSpawn-20210811203113.xml
2021-08-11T20:43:51.0387734Z Generated XML report: test-reports/python-unittest/distributed.rpc.test_tensorpipe_agent/TEST-TensorPipeDdpUnderDistAutogradTestWithSpawn-20210811203113.xml
2021-08-11T20:43:51.0520714Z Generated XML report: test-reports/python-unittest/distributed.rpc.test_tensorpipe_agent/TEST-TensorPipeDistAutogradTestWithSpawn-20210811203113.xml

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.

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. ghstack-source-id: 52d7abb Pull Request resolved: #62024

ezyang · 2021-07-22T14:06:40Z

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

ezyang · 2021-07-22T14:59:21Z

rebasing this stack; @ysiraichi you'll need to ghstack checkout after this

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. ghstack-source-id: 289a1e4 Pull Request resolved: #62024

ezyang · 2021-07-22T15:03:24Z

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

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. ghstack-source-id: f66373c Pull Request resolved: #62024

ezyang · 2021-07-25T04:36:22Z

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

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

ysiraichi · 2021-07-27T08:59:19Z

Rebased.

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. ghstack-source-id: 98c6d92 Pull Request resolved: #62024

ezyang · 2021-07-27T16:43:08Z

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

ezyang · 2021-07-27T16:46:14Z

aten/src/ATen/native/ReduceOps.cpp

+  auto shape = meta::get_reduction_shape(self, {}, false);
+  Tensor result = at::empty(shape, self.options().dtype(dtype));
+  impl_func_prod(self, {}, false, dtype, result);
+  return result;


This looks suspicious. How come we have to define prod here?

The problem is that we can't call prod.dim_int since it expects only one dimension to reduce (prodreduces all of them). This differs from sum.dim_IntList, that expects many dimensions: int[1] (i.e. we can call sum.dim_IntList inside sum kernel)

aten/src/ATen/native/ReduceOps.cpp

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. ghstack-source-id: a865eb8 Pull Request resolved: #62024

ezyang · 2021-07-27T17:14:44Z

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

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

ezyang · 2021-08-04T19:27:49Z

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

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Differential Revision: [D29847122](https://our.internmc.facebook.com/intern/diff/D29847122) [ghstack-poisoned]

ezyang · 2021-08-11T21:03:32Z

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

facebook-github-bot · 2021-08-13T15:28:06Z

@ezyang merged this pull request in 07bb6e4.

Summary: Pull Request resolved: #62024 Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. Test Plan: Imported from OSS Reviewed By: ejguan Differential Revision: D29847122 Pulled By: ezyang fbshipit-source-id: a6680a6cf6e59bb46b8ffe7bf2a3a611d6e0fd14

Port prod kernel to structured kernels.

d5d8359

Tracking issue: #55070 In this PR, I also broke down the meta functions of other reduction kernels (e.g. `all`, `argmax`, `sum`) into the composition of common patterns. [ghstack-poisoned]

This was referenced Jul 22, 2021

Port sum.dim_IntList kernel to structured kernels. #61642

Closed

Port mean kernel to structured kernels. #61643

Closed

facebook-github-bot added the cla signed label Jul 22, 2021

pytorchbot added the open source label Jul 22, 2021

ysiraichi marked this pull request as ready for review July 25, 2021 04:38

ysiraichi requested a review from ezyang as a code owner July 25, 2021 04:38

ysiraichi requested a review from bdhirsh July 25, 2021 04:39

ezyang reviewed Jul 27, 2021

View reviewed changes

aten/src/ATen/native/ReduceOps.cpp Show resolved Hide resolved

ezyang approved these changes Jul 27, 2021

View reviewed changes

ysiraichi mentioned this pull request Aug 4, 2021

Port norm kernel to structured kernels. #62711

Closed

facebook-github-bot closed this in 07bb6e4 Aug 13, 2021

facebook-github-bot added the Merged label Aug 13, 2021

facebook-github-bot deleted the gh/ysiraichi/21/head branch August 17, 2021 14:17

Port prod kernel to structured kernels. #62024

Port prod kernel to structured kernels. #62024

Uh oh!

Conversation

ysiraichi commented Jul 22, 2021 • edited by ezyang Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

facebook-github-bot commented Jul 22, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful links

💊 CI failures summary and remediations

🕵️ 2 new failures recognized by patterns

pytorch_xla_linux_bionic_py3_6_clang9_build (1/2)

pytorch_linux_xenial_py3_6_gcc5_4_build (2/2)

1 failure not recognized by patterns:

❄️ 1 failure tentatively classified as flaky

linux-bionic-py3.8-gcc9-coverage / test (default, 2, 2, linux.2xlarge) (1/1)

Uh oh!

ezyang commented Jul 22, 2021

Uh oh!

ezyang commented Jul 22, 2021

Uh oh!

ezyang commented Jul 22, 2021

Uh oh!

ezyang commented Jul 25, 2021

Uh oh!

ysiraichi commented Jul 27, 2021

Uh oh!

ezyang commented Jul 27, 2021

Uh oh!

ezyang Jul 27, 2021

Choose a reason for hiding this comment

Uh oh!

ysiraichi Jul 28, 2021

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ezyang commented Jul 27, 2021

Uh oh!

ezyang commented Aug 4, 2021

Uh oh!

ezyang commented Aug 11, 2021

Uh oh!

facebook-github-bot commented Aug 13, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Port `prod` kernel to structured kernels. #62024

Port `prod` kernel to structured kernels. #62024

ysiraichi commented Jul 22, 2021 •

edited by ezyang

Loading

facebook-github-bot commented Jul 22, 2021 •

edited

Loading