OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

[CI] Move bazel from 11.3 to 11.6 (#84314)

Browse Source 
In process of doing so have to:
- Delete `/usr/local/cuda-11.6/cuda-11.6` symlink to self, otherwise Bazel builds fail with
```
ERROR: circular symlinks detected
[start of symlink cycle]
/usr/local/cuda-11.6/cuda-11.6
[end of symlink cycle]
```
- Add `-DCUB_WRAPPED_NAMESPACE=at_cuda_detail"` to `COMMON_COPTS` if building with CUDA, to mimic the behaviour in
4b8ae04788/cmake/Dependencies.cmake (L1664-L1668)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/84314
Approved by: https://github.com/ngimel, https://github.com/atalman
This commit is contained in:
Nikita Shulga 2022-08-31 23:02:42 +00:00 committed by PyTorch MergeBot
parent fddfc4488a
commit e8885a872c
3 changed files with 12 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
.circleci/docker/ubuntu-cuda/Dockerfile
View File

@ -122,5 +122,9 @@ COPY ./common/install_cudnn.sh install_cudnn.sh
RUN if [ "${CUDNN_VERSION}" -eq 8 ]; then bash install_cudnn.sh; fi RUN if [ "${CUDNN_VERSION}" -eq 8 ]; then bash install_cudnn.sh; fi
RUN rm install_cudnn.sh RUN rm install_cudnn.sh
# Delete /usr/local/cuda-11.X/cuda-11.X symlinks
RUN if [ -h /usr/local/cuda-11.6/cuda-11.6 ]; then rm /usr/local/cuda-11.6/cuda-11.6; fi
RUN if [ -h /usr/local/cuda-11.7/cuda-11.7 ]; then rm /usr/local/cuda-11.7/cuda-11.7; fi
USER jenkins USER jenkins
CMD ["bash"] CMD ["bash"]

8
.github/workflows/pull.yml vendored
View File

@ -253,12 +253,12 @@ jobs:
cuda-version: "11.6" cuda-version: "11.6"
sync-tag: win-cuda-build sync-tag: win-cuda-build
linux-xenial-cuda11_3-py3_7-gcc7-bazel-test: linux-bionic-cuda11_6-py3_10-gcc7-bazel-test:
name: linux-xenial-cuda11.3-py3.7-gcc7-bazel-test name: linux-bionic-cuda11.6-py3.10-gcc7-bazel-test
uses: ./.github/workflows/_bazel-build-test.yml uses: ./.github/workflows/_bazel-build-test.yml
with: with:
build-environment: linux-xenial-cuda11.3-py3.7-gcc7-bazel-test build-environment: linux-bionicl-cuda11.6-py3.10-gcc7-bazel-test
docker-image-name: pytorch-linux-xenial-cuda11.3-cudnn8-py3-gcc7 docker-image-name: pytorch-linux-bionic-cuda11.6-cudnn8-py3-gcc7
linux-xenial-py3-clang5-android-ndk-r19c-gradle-custom-build-single: linux-xenial-py3-clang5-android-ndk-r19c-gradle-custom-build-single:
name: linux-xenial-py3-clang5-android-ndk-r19c-gradle-custom-build-single name: linux-xenial-py3-clang5-android-ndk-r19c-gradle-custom-build-single

4
BUILD.bazel
View File

@ -28,6 +28,10 @@ COMMON_COPTS = [
] + if_cuda([ ] + if_cuda([
"-DUSE_CUDA", "-DUSE_CUDA",
"-DUSE_CUDNN", "-DUSE_CUDNN",
# TODO: This should be passed only when building for CUDA-11.5 or newer
# use cub in a safe manner, see:
# https://github.com/pytorch/pytorch/pull/55292
"-DCUB_WRAPPED_NAMESPACE=at_cuda_detail",
]) ])
aten_generation_srcs = ["aten/src/ATen/native/native_functions.yaml"] + ["aten/src/ATen/native/tags.yaml"] + glob(["aten/src/ATen/templates/**"]) aten_generation_srcs = ["aten/src/ATen/native/native_functions.yaml"] + ["aten/src/ATen/native/tags.yaml"] + glob(["aten/src/ATen/templates/**"])