Skip to content

[BUG] submodule-sync failed to find nvtx3::nvtx3-cpp target in branch-25.04 #2992

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
pxLi opened this issue Feb 27, 2025 · 6 comments
Closed
Labels
bug Something isn't working

Comments

@pxLi
Copy link
Member

pxLi commented Feb 27, 2025

Describe the bug

updates in this run
cudf commit to rapidsai/cudf@e5d866b
rapids-cmake pinned to rapidsai/rapids-cmake@a8288b5

versions.json diff comparing to branch-25.04

     "rmm" :
     {
       "always_download" : true,
       "git_shallow" : false,
-      "git_tag" : "a7577403c90baf6ea93c67e71703933efa884570",
+      "git_tag" : "d1a5f1baf61128ff35222df5e9be7744bec6d66e",
       "git_url" : "https://github.com/rapidsai/rmm.git",
       "version" : "25.04"
     },
05:20:46  [INFO]      [exec] -- Configuring done (156.0s)
05:20:52  [INFO]      [exec] CMake Error at CMakeLists.txt:932 (target_link_libraries):
05:20:52  [INFO]      [exec]   Target "cudf" links to:
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]     nvtx3::nvtx3-cpp
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]   but the target was not found.  Possible reasons include:
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]     * There is a typo in the target name.
05:20:52  [INFO]      [exec]     * A find_package call is missing for an IMPORTED target.
05:20:52  [INFO]      [exec]     * An ALIAS target is missing.
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] CMake Error at /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4412/target/libcudf/cmake-build/_deps/rmm-src/CMakeLists.txt:103 (target_link_libraries):
05:20:52  [INFO]      [exec]   The link interface of target "rmm" contains:
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]     nvtx3::nvtx3-cpp
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]   but the target was not found.  Possible reasons include:
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]     * There is a typo in the target name.
05:20:52  [INFO]      [exec]     * A find_package call is missing for an IMPORTED target.
05:20:52  [INFO]      [exec]     * An ALIAS target is missing.
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] -- Generating done (5.1s)
05:20:52  [INFO]      [exec] CMake Warning:
05:20:52  [INFO]      [exec]   Manually-specified variables were not used by the project:
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec]     CUDA_STATIC_CUFILE
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] 
05:20:52  [INFO]      [exec] CMake Generate step failed.  Build files cannot be regenerated correctly.

Steps/Code to reproduce bug
Please provide a list of steps or a code sample to reproduce the issue.
Avoid posting private or sensitive data.

Expected behavior
A clear and concise description of what you expected to happen.

Environment details (please complete the following information)

  • Environment location: [Standalone, YARN, Kubernetes, Cloud(specify cloud provider)]
  • Spark configuration settings related to the issue

Additional context
Add any other context about the problem here.

@pxLi pxLi added ? - Needs Triage bug Something isn't working labels Feb 27, 2025
@pxLi pxLi changed the title [BUG] submodule-sync failed to find nvtx3::nvtx3-cpp target [BUG] submodule-sync failed to find nvtx3::nvtx3-cpp target in branch-25.04 Feb 27, 2025
@pxLi
Copy link
Member Author

pxLi commented Feb 27, 2025

confirmed this is related to the rapids-cmake update rapidsai/rapids-cmake#753, revert this commit in thirdparty/cudf-pins/rapids-cmake.sha to 17f21fad85c9de15040f485a385d20da12207dcf could pass JNI build with other updates in pinned mode

@robertmaynard do you have any idea why CPM upgrade could lead to this issue? thanks

previous

[2025-02-26T13:12:08.863Z] [INFO]      [exec] -- CPM: Adding package nvtx3@3.1.0 (v3.1.0)
[2025-02-26T13:12:08.864Z] [INFO]      [exec] CMake Warning (dev) at /usr/local/cmake-3.30.4-linux-x86_64/share/cmake-3.30/Modules/FetchContent.cmake:1953 (message):
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   Calling FetchContent_Populate(nvtx3) is deprecated, call
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   FetchContent_MakeAvailable(nvtx3) instead.  Policy CMP0169 can be set to
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   OLD to allow FetchContent_Populate(nvtx3) to be called directly for now,
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   but the ability to call it with declared details will be removed completely
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   in a future version.
[2025-02-26T13:12:08.864Z] [INFO]      [exec] Call Stack (most recent call first):
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4410/target/libcudf/cmake-build/cmake/CPM_0.40.0.cmake:1074 (FetchContent_Populate)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4410/target/libcudf/cmake-build/cmake/CPM_0.40.0.cmake:868 (cpm_fetch_package)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4410/target/libcudf/cmake-build/cmake/CPM_0.40.0.cmake:306 (CPMAddPackage)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4410/target/libcudf/cmake-build/_deps/rapids-cmake-src/rapids-cmake/cpm/find.cmake:189 (CPMFindPackage)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-dev-4410/target/libcudf/cmake-build/_deps/rapids-cmake-src/rapids-cmake/cpm/nvtx3.cmake:75 (rapids_cpm_find)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   cmake/thirdparty/get_nvtx.cmake:21 (rapids_cpm_nvtx3)
[2025-02-26T13:12:08.864Z] [INFO]      [exec]   cmake/thirdparty/get_nvtx.cmake:25 (find_and_configure_nvtx)

with rapidsai/rapids-cmake#753

[2025-02-26T17:12:17.843Z] [INFO]      [exec] -- CPM: Adding package nvtx3@3.1.0 (v3.1.0)
[2025-02-26T17:13:55.459Z] [INFO]      [exec] -- CPM: Adding package nvtx3@3.1.0 (e170594ac7cf1dac584da473d4ca9301087090c1)
[2025-02-26T17:13:55.459Z] [INFO]      [exec] Cloning into 'nvtx3-src'...
[2025-02-26T17:13:57.968Z] [INFO]      [exec] HEAD is now at e170594 Merge branch 'dev' into release-v3
[2025-02-26T17:13:58.897Z] [INFO]      [exec] error: include/fmt/base.h: No such file or directory
[2025-02-26T17:14:35.406Z] [INFO]      [exec] error: cub/cub/thread/thread_sort.cuh: No such file or directory
[2025-02-26T17:14:35.406Z] [INFO]      [exec] error: cub/cub/block/block_merge_sort.cuh: No such file or directory
[2025-02-26T17:14:35.406Z] [INFO]      [exec] error: cub/cub/device/dispatch/tuning/tuning_scan.cuh: No such file or directory
[2025-02-26T17:14:35.406Z] [INFO]      [exec] error: cub/cub/device/dispatch/dispatch_reduce.cuh: No such file or directory
[2025-02-26T17:14:35.406Z] [INFO]      [exec] error: cub/cub/device/dispatch/dispatch_radix_sort.cuh: No such file or directory
[2025-02-26T17:15:00.722Z] [INFO]      [exec] CMake Error at CMakeLists.txt:932 (target_link_libraries):
[2025-02-26T17:15:00.722Z] [INFO]      [exec]   Target "cudf" links to:
[2025-02-26T17:15:00.722Z] [INFO]      [exec] 
[2025-02-26T17:15:00.722Z] [INFO]      [exec]     nvtx3::nvtx3-cpp
....

@robertmaynard
Copy link
Contributor

@pxLi Is it only the nvtx3 module that you are seeing this regression?

The CPM bump changed lots of internal behaviors due to resolving the policy CMP0169 warnings it was previously generating. So it might be due to that, or it could be due to the new minimum required CMake version we now have

@robertmaynard
Copy link
Contributor

I have a local reproducer

@robertmaynard
Copy link
Contributor

robertmaynard commented Feb 28, 2025

I now know what is wrong, and will be starting work on a fix. But given the amount of effort required I expect I will roll the CPM version back in rapids-cmake for 25.04

@robertmaynard
Copy link
Contributor

rapidsai/rapids-cmake#788

@pxLi
Copy link
Member Author

pxLi commented Mar 4, 2025

close as the CPM rollback fixed the issue rapidsai/rapids-cmake#789

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

3 participants