more changes #4
Merged
more changes #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
hetong007
added a commit
that referenced
this pull request
Aug 6, 2020
* PPIDataset * Revert "PPIDataset" This reverts commit 264bd0c. * update data rst * update data doc and docstring * API doc rst for dataset * docstring * update api doc * add url format * update docstring * update citation graph * update knowledge graph * update gc datasets * fix index * Rst fix (#3) * Fix syntax * syntax * update docstring * update doc (#4) * final update * fix rdflib * fix rdf Co-authored-by: HuXiangkun <[email protected]> Co-authored-by: Ubuntu <[email protected]> Co-authored-by: xiang song(charlie.song) <[email protected]>
VoVAllen
referenced
this pull request
in VoVAllen/dgl
Nov 13, 2020
[KVstore] Implement async_pull API
GMNGeoffrey
added a commit
to GMNGeoffrey/dgl
that referenced
this pull request
Jan 29, 2025
There are definitely still runtime issues, but this handles all of the compilation failures (requires clang-19 and bleeding-edge ROCM). Mostly straightforward. The ones that aren't. - The AtomicFPOp was adapated from [PyTorch](https://github.com/pytorch/pytorch/blob/bef103934a25d848838a7642a8d6a2f523e7dfff/aten/src/ATen/cuda/Atomic.cuh#L39). - The handling of legacy cusparse is something I messed up originally. The various CUDA version checks need to be in sync because the legacy version creates a transposed output that then needs to be flipped back, so I factored out a shared macro. - There is something weird where HIP doesn't like the logging function and thinks a device function is calling a host function. I was never able to get it to work, but it doesn't seem hugely important, so I think it's ok to punt for now. A couple of these changes might not be strictly necessary to make the build work (like adding `__HIP_DEVICE_COMPILE__` to some of the `__CUDA_ARCH__` checks) because I grabbed changes from my fully working draft by file and just reverted the obviously more complicated ones. It didn't seem worth reverting these uncomplicated ones. Specifically, I'm using this build configuration (via a CMake preset, but dumped as a command line): ```shell ROCM_PATH="${HOME}/rocm/opt/rocm" \ CC=/usr/bin/clang-19 \ CXX=/usr/bin/clang++-19 \ /usr/local/bin/cmake \ -DUSE_ROCM=ON \ -DCMAKE_PREFIX_PATH=/home/gcmn/rocm/opt/rocm \ -DCMAKE_POSITION_INDEPENDENT_CODE=ON \ -DROCM_WARN_TOOLCHAIN_VAR=OFF \ -DCMAKE_EXPORT_COMPILE_COMMANDS=ON \ -DCMAKE_INSTALL_PREFIX=/home/gcmn/src/dgl/out/install/rocm \ -DCMAKE_C_COMPILER=/usr/bin/clang-19 \ -DCMAKE_CXX_COMPILER=/usr/bin/clang++-19 \ -DCMAKE_C_COMPILER_LAUNCHER=ccache \ -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \ -DCMAKE_BUILD_TYPE=Debug \ -DBUILD_TYPE=dev \ -DBUILD_CPP_TEST=ON \ -DBUILD_GRAPHBOLT=OFF \ -DBUILD_SPARSE=ON \ -DUSE_LIBXSMM=OFF \ -DUSE_OPENMP=ON \ -DPython3_FIND_VIRTUALENV=ONLY \ -S/home/gcmn/src/dgl \ -B/home/gcmn/src/dgl/out/build/rocm \ -G Ninja ```
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
No description provided.