Skip to content

Add static_runtime::fused_equally_split #66881

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.

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

Closed
wants to merge 1 commit into from

Conversation

pilleye
Copy link
Member

@pilleye pilleye commented Oct 19, 2021

Summary: Adds static_runtime::fused_equally_split operator and removes is_fused logic from original operator. Modifies FuseUnpackListV2 to map fb::equally_split to this new operator.

Test Plan:

adityapillai@5960 /data/sandcastle/boxes/fbsource/fbcode 1m 13s
❯ buck test //caffe2/benchmarks/static_runtime/fb:test_fb_operators

and sandcastle
strange_what_could_go_wrong

Differential Revision: D31742293

@pytorch-probot
Copy link

pytorch-probot bot commented Oct 19, 2021

CI Flow Status

⚛️ CI Flow

Ruleset - Version: v1
Ruleset - File: https://github.com/pilleye/pytorch/blob/717f0d977530cde9ec7e062911ccef11692a066f/.github/generated-ciflow-ruleset.json
PR ciflow labels: ciflow/default

Workflows Labels (bold enabled) Status
Triggered Workflows
linux-bionic-py3.6-clang9 ciflow/all, ciflow/cpu, ciflow/default, ciflow/linux, ciflow/noarch, ciflow/xla ✅ triggered
linux-vulkan-bionic-py3.6-clang9 ciflow/all, ciflow/cpu, ciflow/default, ciflow/linux, ciflow/vulkan ✅ triggered
linux-xenial-cuda11.3-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/default, ciflow/linux ✅ triggered
linux-xenial-py3-clang5-mobile-build ciflow/all, ciflow/default, ciflow/linux, ciflow/mobile ✅ triggered
linux-xenial-py3-clang5-mobile-custom-build-dynamic ciflow/all, ciflow/default, ciflow/linux, ciflow/mobile ✅ triggered
linux-xenial-py3-clang5-mobile-custom-build-static ciflow/all, ciflow/default, ciflow/linux, ciflow/mobile ✅ triggered
linux-xenial-py3.6-clang7-asan ciflow/all, ciflow/cpu, ciflow/default, ciflow/linux, ciflow/sanitizers ✅ triggered
linux-xenial-py3.6-clang7-onnx ciflow/all, ciflow/cpu, ciflow/default, ciflow/linux, ciflow/onnx ✅ triggered
linux-xenial-py3.6-gcc5.4 ciflow/all, ciflow/cpu, ciflow/default, ciflow/linux ✅ triggered
linux-xenial-py3.6-gcc7-bazel-test ciflow/all, ciflow/bazel, ciflow/cpu, ciflow/default, ciflow/linux ✅ triggered
win-vs2019-cpu-py3 ciflow/all, ciflow/cpu, ciflow/default, ciflow/win ✅ triggered
win-vs2019-cuda11.3-py3 ciflow/all, ciflow/cuda, ciflow/default, ciflow/win ✅ triggered
Skipped Workflows
libtorch-linux-xenial-cuda10.2-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/libtorch, ciflow/linux 🚫 skipped
libtorch-linux-xenial-cuda11.3-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/libtorch, ciflow/linux 🚫 skipped
linux-bionic-cuda10.2-py3.9-gcc7 ciflow/all, ciflow/cuda, ciflow/linux, ciflow/slow 🚫 skipped
linux-xenial-cuda10.2-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/linux, ciflow/slow 🚫 skipped
linux-xenial-py3-clang5-mobile-code-analysis ciflow/all, ciflow/linux, ciflow/mobile 🚫 skipped
parallelnative-linux-xenial-py3.6-gcc5.4 ciflow/all, ciflow/cpu, ciflow/linux 🚫 skipped
periodic-libtorch-linux-xenial-cuda11.1-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/libtorch, ciflow/linux, ciflow/scheduled 🚫 skipped
periodic-linux-xenial-cuda10.2-py3-gcc7-slow-gradcheck ciflow/all, ciflow/cuda, ciflow/linux, ciflow/scheduled, ciflow/slow, ciflow/slow-gradcheck 🚫 skipped
periodic-linux-xenial-cuda11.1-py3.6-gcc7 ciflow/all, ciflow/cuda, ciflow/linux, ciflow/scheduled 🚫 skipped
periodic-win-vs2019-cuda11.1-py3 ciflow/all, ciflow/cuda, ciflow/scheduled, ciflow/win 🚫 skipped
puretorch-linux-xenial-py3.6-gcc5.4 ciflow/all, ciflow/cpu, ciflow/linux 🚫 skipped

You can add a comment to the PR and tag @pytorchbot with the following commands:
# ciflow rerun, "ciflow/default" will always be added automatically
@pytorchbot ciflow rerun

# ciflow rerun with additional labels "-l <ciflow/label_name>", which is equivalent to adding these labels manually and trigger the rerun
@pytorchbot ciflow rerun -l ciflow/scheduled -l ciflow/slow

For more information, please take a look at the CI Flow Wiki.

@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented Oct 19, 2021

🔗 Helpful links

💊 CI failures summary and remediations

As of commit 717f0d9 (more details on the Dr. CI page):


  • 6/6 failures possibly* introduced in this PR
    • 1/6 non-scanned failure(s)

🕵️ 5 new failures recognized by patterns

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

See GitHub Actions build linux-bionic-py3.6-clang9 / test (noarch, 1, 1, linux.2xlarge) (1/5)

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

2021-10-21T22:17:04.4004321Z FAIL [0.006s]: tes...v_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:16:52.9268478Z   test_torchvision_models_vgg16_bn (__main__.TestVisionTracing) ... ok (1.949s)
2021-10-21T22:16:54.8779140Z   test_torchvision_models_vgg19 (__main__.TestVisionTracing) ... ok (1.951s)
2021-10-21T22:16:56.9470752Z   test_torchvision_models_vgg19_bn (__main__.TestVisionTracing) ... ok (2.069s)
2021-10-21T22:16:57.7906782Z   test_torchvision_models_video_mc3_18 (__main__.TestVisionTracing) ... ok (0.844s)
2021-10-21T22:16:58.9679652Z   test_torchvision_models_video_r2plus1d_18 (__main__.TestVisionTracing) ... ok (1.177s)
2021-10-21T22:17:00.1442991Z   test_torchvision_models_video_r3d_18 (__main__.TestVisionTracing) ... ok (1.176s)
2021-10-21T22:17:02.8683239Z   test_torchvision_models_wide_resnet101_2 (__main__.TestVisionTracing) ... ok (2.724s)
2021-10-21T22:17:04.3999895Z   test_torchvision_models_wide_resnet50_2 (__main__.TestVisionTracing) ... ok (1.531s)
2021-10-21T22:17:04.4002335Z 
2021-10-21T22:17:04.4003167Z ======================================================================
2021-10-21T22:17:04.4004321Z FAIL [0.006s]: test_get_torch_func_signature_exhaustive_cov_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:17:04.4006234Z ----------------------------------------------------------------------
2021-10-21T22:17:04.4007090Z Traceback (most recent call last):
2021-10-21T22:17:04.4007987Z   File "test_fx.py", line 3254, in test_get_torch_func_signature_exhaustive
2021-10-21T22:17:04.4008857Z     op(*bound_args.args, **bound_args.kwargs)
2021-10-21T22:17:04.4010406Z   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_methods_invocations.py", line 666, in __call__
2021-10-21T22:17:04.4011536Z     return self.op(*args, **kwargs)
2021-10-21T22:17:04.4012514Z RuntimeError: cov(): weights sum to zero, can't be normalized
2021-10-21T22:17:04.4013104Z 
2021-10-21T22:17:04.4013903Z During handling of the above exception, another exception occurred:
2021-10-21T22:17:04.4014572Z 

See GitHub Actions build linux-bionic-py3.6-clang9 / test (default, 1, 2, linux.2xlarge) (2/5)

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

2021-10-21T22:19:25.2529313Z FAIL [0.006s]: tes...v_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:19:13.7634099Z   test_torchvision_models_vgg16_bn (__main__.TestVisionTracing) ... ok (1.847s)
2021-10-21T22:19:15.6904470Z   test_torchvision_models_vgg19 (__main__.TestVisionTracing) ... ok (1.927s)
2021-10-21T22:19:17.7956935Z   test_torchvision_models_vgg19_bn (__main__.TestVisionTracing) ... ok (2.105s)
2021-10-21T22:19:18.6714168Z   test_torchvision_models_video_mc3_18 (__main__.TestVisionTracing) ... ok (0.876s)
2021-10-21T22:19:19.8268808Z   test_torchvision_models_video_r2plus1d_18 (__main__.TestVisionTracing) ... ok (1.155s)
2021-10-21T22:19:20.8900723Z   test_torchvision_models_video_r3d_18 (__main__.TestVisionTracing) ... ok (1.063s)
2021-10-21T22:19:23.6704096Z   test_torchvision_models_wide_resnet101_2 (__main__.TestVisionTracing) ... ok (2.780s)
2021-10-21T22:19:25.2526763Z   test_torchvision_models_wide_resnet50_2 (__main__.TestVisionTracing) ... ok (1.582s)
2021-10-21T22:19:25.2527550Z 
2021-10-21T22:19:25.2528072Z ======================================================================
2021-10-21T22:19:25.2529313Z FAIL [0.006s]: test_get_torch_func_signature_exhaustive_cov_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:19:25.2531200Z ----------------------------------------------------------------------
2021-10-21T22:19:25.2532028Z Traceback (most recent call last):
2021-10-21T22:19:25.2533094Z   File "test_fx.py", line 3254, in test_get_torch_func_signature_exhaustive
2021-10-21T22:19:25.2534060Z     op(*bound_args.args, **bound_args.kwargs)
2021-10-21T22:19:25.2535612Z   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_methods_invocations.py", line 666, in __call__
2021-10-21T22:19:25.2536789Z     return self.op(*args, **kwargs)
2021-10-21T22:19:25.2537826Z RuntimeError: cov(): weights sum to zero, can't be normalized
2021-10-21T22:19:25.2538425Z 
2021-10-21T22:19:25.2539208Z During handling of the above exception, another exception occurred:
2021-10-21T22:19:25.2539888Z 

See GitHub Actions build linux-xenial-py3.6-clang7-asan / test (default, 2, 2, linux.2xlarge) (3/5)

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

2021-10-21T23:11:28.3982986Z FAIL [0.016s]: tes...v_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T23:10:45.0077251Z   test_torchvision_models_vgg16_bn (__main__.TestVisionTracing) ... ok (7.648s)
2021-10-21T23:10:52.8898386Z   test_torchvision_models_vgg19 (__main__.TestVisionTracing) ... ok (7.882s)
2021-10-21T23:11:01.5641098Z   test_torchvision_models_vgg19_bn (__main__.TestVisionTracing) ... ok (8.674s)
2021-10-21T23:11:04.7497765Z   test_torchvision_models_video_mc3_18 (__main__.TestVisionTracing) ... ok (3.186s)
2021-10-21T23:11:09.1079162Z   test_torchvision_models_video_r2plus1d_18 (__main__.TestVisionTracing) ... ok (4.358s)
2021-10-21T23:11:12.9949837Z   test_torchvision_models_video_r3d_18 (__main__.TestVisionTracing) ... ok (3.887s)
2021-10-21T23:11:22.9397317Z   test_torchvision_models_wide_resnet101_2 (__main__.TestVisionTracing) ... ok (9.945s)
2021-10-21T23:11:28.3979384Z   test_torchvision_models_wide_resnet50_2 (__main__.TestVisionTracing) ... ok (5.458s)
2021-10-21T23:11:28.3980120Z 
2021-10-21T23:11:28.3980654Z ======================================================================
2021-10-21T23:11:28.3982986Z FAIL [0.016s]: test_get_torch_func_signature_exhaustive_cov_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T23:11:28.3984316Z ----------------------------------------------------------------------
2021-10-21T23:11:28.3984860Z Traceback (most recent call last):
2021-10-21T23:11:28.3985720Z   File "test_fx.py", line 3254, in test_get_torch_func_signature_exhaustive
2021-10-21T23:11:28.3986411Z     op(*bound_args.args, **bound_args.kwargs)
2021-10-21T23:11:28.3987664Z   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_methods_invocations.py", line 666, in __call__
2021-10-21T23:11:28.3988726Z     return self.op(*args, **kwargs)
2021-10-21T23:11:28.3989789Z RuntimeError: cov(): weights sum to zero, can't be normalized
2021-10-21T23:11:28.3990185Z 
2021-10-21T23:11:28.3990987Z During handling of the above exception, another exception occurred:
2021-10-21T23:11:28.3991649Z 

See GitHub Actions build linux-xenial-py3.6-gcc5.4 / test (default, 1, 2, linux.2xlarge) (4/5)

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

2021-10-21T22:08:41.3205732Z FAIL [0.006s]: tes...v_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:08:29.2153251Z   test_torchvision_models_vgg16_bn (__main__.TestVisionTracing) ... ok (1.991s)
2021-10-21T22:08:31.2964289Z   test_torchvision_models_vgg19 (__main__.TestVisionTracing) ... ok (2.081s)
2021-10-21T22:08:33.5440241Z   test_torchvision_models_vgg19_bn (__main__.TestVisionTracing) ... ok (2.248s)
2021-10-21T22:08:34.4242613Z   test_torchvision_models_video_mc3_18 (__main__.TestVisionTracing) ... ok (0.880s)
2021-10-21T22:08:35.6280042Z   test_torchvision_models_video_r2plus1d_18 (__main__.TestVisionTracing) ... ok (1.204s)
2021-10-21T22:08:36.7388943Z   test_torchvision_models_video_r3d_18 (__main__.TestVisionTracing) ... ok (1.111s)
2021-10-21T22:08:39.6717359Z   test_torchvision_models_wide_resnet101_2 (__main__.TestVisionTracing) ... ok (2.933s)
2021-10-21T22:08:41.3203712Z   test_torchvision_models_wide_resnet50_2 (__main__.TestVisionTracing) ... ok (1.648s)
2021-10-21T22:08:41.3204346Z 
2021-10-21T22:08:41.3204725Z ======================================================================
2021-10-21T22:08:41.3205732Z FAIL [0.006s]: test_get_torch_func_signature_exhaustive_cov_cpu_float32 (__main__.TestOperatorSignaturesCPU)
2021-10-21T22:08:41.3207414Z ----------------------------------------------------------------------
2021-10-21T22:08:41.3208285Z Traceback (most recent call last):
2021-10-21T22:08:41.3208800Z   File "test_fx.py", line 3254, in test_get_torch_func_signature_exhaustive
2021-10-21T22:08:41.3209319Z     op(*bound_args.args, **bound_args.kwargs)
2021-10-21T22:08:41.3210192Z   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_methods_invocations.py", line 666, in __call__
2021-10-21T22:08:41.3210841Z     return self.op(*args, **kwargs)
2021-10-21T22:08:41.3211437Z RuntimeError: cov(): weights sum to zero, can't be normalized
2021-10-21T22:08:41.3211769Z 
2021-10-21T22:08:41.3212200Z During handling of the above exception, another exception occurred:
2021-10-21T22:08:41.3212570Z 

See CircleCI build pytorch_linux_xenial_py3_6_gcc5_4_test (5/5)

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

Oct 21 22:03:42 FAIL [0.007s]: test_get_torch_f...v_cpu_float32 (__main__.TestOperatorSignaturesCPU)
Oct 21 22:03:26   test_torchvision_models_vgg16_bn (__main__.TestVisionTracing) ... ok (2.501s)
Oct 21 22:03:29   test_torchvision_models_vgg19 (__main__.TestVisionTracing) ... ok (2.619s)
Oct 21 22:03:31   test_torchvision_models_vgg19_bn (__main__.TestVisionTracing) ... ok (2.713s)
Oct 21 22:03:33   test_torchvision_models_video_mc3_18 (__main__.TestVisionTracing) ... ok (1.402s)
Oct 21 22:03:34   test_torchvision_models_video_r2plus1d_18 (__main__.TestVisionTracing) ... ok (1.612s)
Oct 21 22:03:36   test_torchvision_models_video_r3d_18 (__main__.TestVisionTracing) ... ok (1.634s)
Oct 21 22:03:40   test_torchvision_models_wide_resnet101_2 (__main__.TestVisionTracing) ... ok (3.789s)
Oct 21 22:03:42   test_torchvision_models_wide_resnet50_2 (__main__.TestVisionTracing) ... ok (2.039s)
Oct 21 22:03:42 
Oct 21 22:03:42 ======================================================================
Oct 21 22:03:42 FAIL [0.007s]: test_get_torch_func_signature_exhaustive_cov_cpu_float32 (__main__.TestOperatorSignaturesCPU)
Oct 21 22:03:42 ----------------------------------------------------------------------
Oct 21 22:03:42 Traceback (most recent call last):
Oct 21 22:03:42   File "test_fx.py", line 3254, in test_get_torch_func_signature_exhaustive
Oct 21 22:03:42     op(*bound_args.args, **bound_args.kwargs)
Oct 21 22:03:42   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_methods_invocations.py", line 666, in __call__
Oct 21 22:03:42     return self.op(*args, **kwargs)
Oct 21 22:03:42 RuntimeError: cov(): weights sum to zero, can't be normalized
Oct 21 22:03:42 
Oct 21 22:03:42 During handling of the above exception, another exception occurred:
Oct 21 22:03:42 

ci.pytorch.org: 1 failed


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.

@facebook-github-bot facebook-github-bot added oncall: jit Add this issue/PR to JIT oncall triage queue fb-exported labels Oct 19, 2021
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D31742293

1 similar comment
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D31742293

@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D31742293

@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D31742293

Summary:
Pull Request resolved: pytorch/pytorch-canary#2

Pull Request resolved: pytorch#66881

Adds `static_runtime::fused_equally_split` operator and removes `is_fused` logic from original operator. Modifies `FuseUnpackListV2` to map `fb::equally_split` to this new operator.

Test Plan:
```
adityapillai@5960 /data/sandcastle/boxes/fbsource/fbcode 1m 13s
❯ buck test //caffe2/benchmarks/static_runtime/fb:test_fb_operators
```
and sandcastle
strange_what_could_go_wrong

Reviewed By: mikeiovine

Differential Revision: D31742293

fbshipit-source-id: a8ee814e82742a14fe8466ecbde3a70de063b87e
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D31742293

@facebook-github-bot
Copy link
Contributor

This pull request has been merged in 40a8a50.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
cla signed fb-exported Merged oncall: jit Add this issue/PR to JIT oncall triage queue
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants