adding Coderabbit - #5638
Conversation
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
seunghwak
left a comment
There was a problem hiding this comment.
It seems like the C++ requirements include many cudf specific items. We need to delete them. Otherwise, looks like a good starting point.
| - Temporary memory not using `cugraph::get_current_device_resource_ref()` | ||
| - Using `thrust::device_vector` or `rmm::device_scalar<T>` instead of `rmm::device_uvector` or `cugraph::detail::device_scalar<T>` | ||
| - Using raw `cudaMemcpyAsync` instead of `cugraph::detail::cuda_memcpy_async` / `memcpy_async` / `memcpy_batch_async` | ||
| - Using `cugraph::detail::make_host_vector{,_async}` instead of `cugraph::detail::make_pinned_vector{,_async}` for small H2D/D2H transfers |
There was a problem hiding this comment.
This also sounds like a cudf specific requirement.
There was a problem hiding this comment.
We don't have make_host_vector or make_pinned_vector.
seunghwak
left a comment
There was a problem hiding this comment.
We don't have
cugraph::detail::make_host_vector{,_async} or cugraph::detail::make_pinned_vector{,_async}, so the line mentioning these functions is irrelevant.
Otherwise, LGTM.
| - Temporary memory not using `cugraph::get_current_device_resource_ref()` | ||
| - Using `thrust::device_vector` or `rmm::device_scalar<T>` instead of `rmm::device_uvector` or `cugraph::detail::device_scalar<T>` | ||
| - Using raw `cudaMemcpyAsync` instead of `cugraph::detail::cuda_memcpy_async` / `memcpy_async` / `memcpy_batch_async` | ||
| - Using `cugraph::detail::make_host_vector{,_async}` instead of `cugraph::detail::make_pinned_vector{,_async}` for small H2D/D2H transfers |
There was a problem hiding this comment.
We don't have make_host_vector or make_pinned_vector.
added the config files for running Coderabbit