KEMBAR78
Remove getScriptRemoteCallType by lw · Pull Request #57854 · pytorch/pytorch · GitHub
Skip to content

Conversation

@lw
Copy link
Contributor

@lw lw commented May 7, 2021

Stack from ghstack:

Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: D28253843

Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
@facebook-github-bot facebook-github-bot added oncall: distributed Add this issue/PR to distributed oncall triage queue cla signed labels May 7, 2021
@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented May 7, 2021

💊 CI failures summary and remediations

As of commit c21688b (more details on the Dr. CI page):


  • 1/1 failures introduced in this PR

🕵️ 1 new failure recognized by patterns

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

See CircleCI build pytorch_xla_linux_bionic_py3_6_clang9_build (1/1)

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

May 21 12:25:12 torch_xla/csrc/aten_xla_type.cp... match any declaration in 'torch_xla::AtenXlaType'
May 21 12:24:57 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/aten_xla_type_default.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/aten_xla_type_default.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:24:59 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/aten_xla_type.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/aten_xla_type.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:25:03 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/batch_norm.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/batch_norm.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:25:11 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/reduction.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/reduction.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:25:12 torch_xla/csrc/aten_xla_type.cpp:1238:25: error: out-of-line definition of 'div' does not match any declaration in 'torch_xla::AtenXlaType'
May 21 12:25:12 at::Tensor AtenXlaType::div(const at::Tensor& self, const at::Tensor& other,
May 21 12:25:12                         ^~~
May 21 12:25:12 /var/lib/jenkins/workspace/xla/torch_xla/csrc/aten_xla_type.h:251:74: note: type of 3rd parameter of member declaration does not match definition ('optional<c10::string_view>' vs 'optional<std::string>')
May 21 12:25:12 static at::Tensor div(const at::Tensor & self, const at::Tensor & other, c10::optional<c10::string_view> rounding_mode);
May 21 12:25:12                                                                          ^
May 21 12:25:12 torch_xla/csrc/aten_xla_type.cpp:1257:26: error: out-of-line definition of 'div_' does not match any declaration in 'torch_xla::AtenXlaType'
May 21 12:25:12 at::Tensor& AtenXlaType::div_(at::Tensor& self, const at::Tensor& other,
May 21 12:25:12                          ^~~~
May 21 12:25:12 /var/lib/jenkins/workspace/xla/torch_xla/csrc/aten_xla_type.h:30:71: note: type of 3rd parameter of member declaration does not match definition ('optional<c10::string_view>' vs 'optional<std::string>')
May 21 12:25:12 static at::Tensor & div_(at::Tensor & self, const at::Tensor & other, c10::optional<c10::string_view> rounding_mode);
May 21 12:25:12                                                                       ^
May 21 12:25:16 2 errors generated.
May 21 12:25:16 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/matrix.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/matrix.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:25:24 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/pooling.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/pooling.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1
May 21 12:25:26 1 warning generated.
May 21 12:25:26 clang-9 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/var/lib/jenkins/workspace/xla -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-bin -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/protobuf_archive/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_protobuf/src -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/eigen_archive -I/var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/external/com_google_absl -I/var/lib/jenkins/workspace -I/var/lib/jenkins/workspace/torch/csrc -I/var/lib/jenkins/workspace/torch/lib/tmp_install/include -I/opt/conda/lib/python3.6/site-packages/torch/include -I/opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -I/opt/conda/lib/python3.6/site-packages/torch/include/TH -I/opt/conda/lib/python3.6/site-packages/torch/include/THC -I/opt/conda/include/python3.6m -c torch_xla/csrc/aten_xla_bridge.cpp -o build/temp.linux-x86_64-3.6/torch_xla/csrc/aten_xla_bridge.o -std=c++14 -Wno-sign-compare -Wno-deprecated-declarations -Wno-return-type -Wno-macro-redefined -Wno-return-std-move -DNDEBUG -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="_clang" -DPYBIND11_STDLIB="_libstdcpp" -DPYBIND11_BUILD_ABI="_cxxabi1002" -DTORCH_EXTENSION_NAME=_XLAC -D_GLIBCXX_USE_CXX11_ABI=1

XLA failure

Job pytorch_xla_linux_bionic_py3_6_clang9_build is failing. Please create an issue with title prefixed by [PT_BREAK] in pytorch/xla and link to to this PR. If you have questions, please reach out to @ailzhang / @dlibenzi / @JackCaoG.


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.

lw added 2 commits May 17, 2021 03:53
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
Copy link
Contributor

@mrshenli mrshenli left a comment

Choose a reason for hiding this comment

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

LGTM

.toRRef());
} else {
ownerRRef = ctx.getOrCreateOwnerRRef(rrefId, returnType);
ownerRRef = ctx.getOrCreateOwnerRRef(rrefId, jitFuture->elementType());
Copy link
Contributor

Choose a reason for hiding this comment

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

I assume runJitFunction and runJitOperator can retrieve type correctly, so that we can directly the the elementType from this Future?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I did look into how they determine the type for the Future that they return, and they seem to be doing some kind of introspection and type inference. Hence to me they seem to be doing the right thing, but I'm not an expert on JIT so I may perhaps miss some nuances?

@lw
Copy link
Contributor Author

lw commented May 18, 2021

There turned out to be test failures because the inferred types were now "too" precise. In particular it seems before we were just inferring a "base" tensor whereas now we infer a "specialized" tensor. I had to thus relax a check later on to accept tensor subclasses.

lw added 3 commits May 18, 2021 02:35
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)

[ghstack-poisoned]
dgl-intel pushed a commit to dgl-intel/pytorch that referenced this pull request May 21, 2021
Pull Request resolved: pytorch#57854

Because OwnerRRefs used to be created before their value was computed, we had to figure out their type ahead of time. After the previous diff, we inverted the order of operations, and we can now first compute the result and then create the OwnerRRef. Which means we can just inspect the value to get its type. Much simpler, and much less likely to get it wrong.
ghstack-source-id: 129567060

Differential Revision: [D28253843](https://our.internmc.facebook.com/intern/diff/D28253843/)
@facebook-github-bot
Copy link
Contributor

This pull request has been merged in 20d02cb.

@facebook-github-bot facebook-github-bot deleted the gh/lw/168/head branch May 25, 2021 14:16
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

cla signed Merged oncall: distributed Add this issue/PR to distributed oncall triage queue

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants