Remove use of CUDA driver API calls from libcudf #11370
Merged
Conversation
This file contains 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
github-actions
bot
added
CMake
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## branch-22.10 #11370 +/- ##
===============================================
Coverage ? 86.41%
===============================================
Files ? 145
Lines ? 22993
Branches ? 0
===============================================
Hits ? 19870
Misses ? 3123
Partials ? 0 Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
This was referenced Aug 19, 2022
rapids-bot bot
pushed a commit
that referenced
this pull request
Aug 25, 2022
This PR enables using [upstream jitify2](https://github.com/NVIDIA/jitify/tree/jitify2) rather than RAPIDS' fork of [jitify2](https://github.com/rapidsai/jitify/tree/cudf_0.19). This enables us to take advantage of the latest additions/improvements to jitify. Most notably: upstream jitify2 dlsym/dlopens `libcuda.so` which enables us to [drop our shared library dependency on `libcuda.so`](#11370). --- Two major issues came up when making the switch: 1. NVIDIA/jitify#107 - I used the workaround mentioned in that issue. Hopefully it is fixed soon and we can eliminate the workaround. 2. We need to pass `-D_FILE_OFFSET_BITS=64` to jitify. Due to limitations in the way conda-forge builds glibc, we must explicitly state we require 64bit file offset support. Authors: - Ashwin Srinath (https://github.com/shwina) Approvers: - Yunsong Wang (https://github.com/PointKernel) - Bradley Dice (https://github.com/bdice) - Robert Maynard (https://github.com/robertmaynard) URL: #11287
shwina
added
improvement
shwina
changed the title
shwina
commented
Aug 29, 2022
davidwendt
requested changes
Aug 29, 2022
robertmaynard
approved these changes
Aug 29, 2022
jrhemstad
reviewed
Aug 29, 2022
davidwendt
approved these changes
Aug 29, 2022
jrhemstad
approved these changes
Aug 29, 2022
magnatelee
approved these changes
Aug 29, 2022
shwina
commented
Aug 29, 2022
shwina
changed the title
|
@gpucibot merge |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Description
This PR adds the final bits that remove
libcuda.soas a shared library dependency oflibcudf. Currently, that's just one place that uses the driver APIcuContextGetCurrent()to get the current CUDA context and maintain a per-context singleton. It was determined in the discussions within this PR that that could be replaced with a per-device singleton and a call to the runtime APIcudaGetDevice()instead.Checklist