[Intel GPU] Fix Windows linkage issue due to invisible structured kernel symbols #137794
+23
−4
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/137794
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit 9361b5a with merge base 41977a0 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
EikanWang
added a commit
that referenced
this pull request
Oct 11, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: ef0ca21 Pull Request resolved: #137794
EikanWang
changed the title
EikanWang
changed the title
EikanWang
changed the title
EikanWang
added
the
triaged
EikanWang
changed the title
…e structured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. And the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define` CAFFE2_BUILD_MAIN_LIB` macro to enable it. [ghstack-poisoned]
…e structured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. And the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define` CAFFE2_BUILD_MAIN_LIB` macro to enable it. [ghstack-poisoned]
…e structured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. Currently, the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. However, we cannot enable `CAFFE2_BUILD_MAIN_LIB` for the Intel GPU ATen library naively. Because the macro not only serves for the `TORCH_API` semantic. It means that the semantic of `TORCH_API` is symbol `hidden`. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Therefore, we need to use ` TORCH_XPU_API` to decorate the produced structure kernels. [ghstack-poisoned]
EikanWang
added a commit
that referenced
this pull request
Oct 12, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: c153216 Pull Request resolved: #137794
EikanWang
added a commit
that referenced
this pull request
Oct 12, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: 3628a5e Pull Request resolved: #137794
…e structured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. Currently, the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. However, we cannot enable `CAFFE2_BUILD_MAIN_LIB` for the Intel GPU ATen library naively. Because the macro not only serves for the `TORCH_API` semantic. It means that the semantic of `TORCH_API` is symbol `hidden`. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Therefore, we need to use ` TORCH_XPU_API` to decorate the produced structure kernels. [ghstack-poisoned]
EikanWang
added a commit
that referenced
this pull request
Oct 12, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: 8714c7c Pull Request resolved: #137794
EikanWang
changed the title
…uctured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. Currently, the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. However, we cannot enable `CAFFE2_BUILD_MAIN_LIB` for the Intel GPU ATen library naively. Because the macro not only serves for the `TORCH_API` semantic. It means that the semantic of `TORCH_API` is symbol `hidden`. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Therefore, we need to use ` TORCH_XPU_API` to decorate the produced structure kernels. [ghstack-poisoned]
EikanWang
added a commit
that referenced
this pull request
Oct 14, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: 8714c7c Pull Request resolved: #137794
EikanWang
commented
Oct 14, 2024
Comment on lines
359
to
360
| # not implemented for 'Boolean' | ||
| "nn.functiona.unfold": {b8}, |
There was a problem hiding this comment.
The latest torch-xpu-ops has supported boolean for unfolad. Therefore, we update this case for the sake of CI.
…uctured kernel symbols" Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. Currently, the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. However, we cannot enable `CAFFE2_BUILD_MAIN_LIB` for the Intel GPU ATen library naively. Because the macro not only serves for the `TORCH_API` semantic. It means that the semantic of `TORCH_API` is symbol `hidden`. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Therefore, we need to use ` TORCH_XPU_API` to decorate the produced structure kernels. cc voznesenskym penguinwu jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang [ghstack-poisoned]
EikanWang
added a commit
that referenced
this pull request
Oct 14, 2024
…nerated kernels are decorated by TORCH_API to control the visibility, while TORCH_API is controled by the CAFFE2_BUILD_MAIN_LIB macro. Intel GPU requires the generated structure kernels to be visible for linking. Therefore, we need to define CAFFE2_BUILD_MAIN_LIB macro to enable it. ghstack-source-id: de0471e Pull Request resolved: #137794
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
jackzhxng
pushed a commit
that referenced
this pull request
Oct 16, 2024
…nel symbols (#137794) Intel GPU aten library(libtorch_xpu) utilizes `torchgen` to generate structure kernels. Currently, the generated structure kernels are decorated by `TORCH_API` to control the visibility, while `TORCH_API` is controlled by the `CAFFE2_BUILD_MAIN_LIB` macro. However, we cannot enable `CAFFE2_BUILD_MAIN_LIB` for the Intel GPU ATen library naively. Because the macro not only serves for the `TORCH_API` semantic. It means that the semantic of `TORCH_API` is symbol `hidden`. https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99 Therefore, we need to use ` TORCH_XPU_API` to decorate the produced structure kernels. Pull Request resolved: #137794 Approved by: https://github.com/atalman ghstack dependencies: #137873
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
Intel GPU aten library(libtorch_xpu) utilizes
torchgento generate structure kernels. Currently, the generated structure kernels are decorated byTORCH_APIto control the visibility, whileTORCH_APIis controlled by theCAFFE2_BUILD_MAIN_LIBmacro. However, we cannot enableCAFFE2_BUILD_MAIN_LIBfor the Intel GPU ATen library naively. Because the macro not only serves for theTORCH_APIsemantic. It means that the semantic ofTORCH_APIis symbolhidden.https://github.com/pytorch/pytorch/blob/main/c10/macros/Export.h#L95-L99
Therefore, we need to use
TORCH_XPU_APIto decorate the produced structure kernels.cc @voznesenskym @penguinwu @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang