Improve cudf::cuda_error #10630
Merged
Improve cudf::cuda_error #10630
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
Signed-off-by: sperlingxx <[email protected]>
sperlingxx
added
cuda
libcudf
|
I don't know how to produce a sticky error in the unit test. |
Codecov Report
@@ Coverage Diff @@
## branch-22.06 #10630 +/- ##
================================================
+ Coverage 86.33% 86.36% +0.02%
================================================
Files 140 142 +2
Lines 22289 22352 +63
================================================
+ Hits 19244 19305 +61
- Misses 3045 3047 +2
Continue to review full report at Codecov.
|
karthikeyann
approved these changes
Apr 11, 2022
jrhemstad
reviewed
Apr 11, 2022
jrhemstad
reviewed
Apr 11, 2022
jrhemstad
reviewed
Apr 11, 2022
jrhemstad
reviewed
Apr 11, 2022
jrhemstad
reviewed
Apr 12, 2022
sperlingxx
added
non-breaking
jrhemstad
reviewed
Apr 13, 2022
| #define CUDF_CUDA_TRY(call) \ | ||
| do { \ | ||
| cudaError_t const status = (call); \ | ||
| if (cudaSuccess != status) { cudf::detail::throw_cuda_error(status, __FILE__, __LINE__); } \ |
There was a problem hiding this comment.
The semicolon after the while(0) should be removed to ensure all uses of the CUDF_CUDA_TRY macro are terminated with a semi-colon.
jrhemstad
reviewed
Apr 13, 2022
jrhemstad
approved these changes
Apr 13, 2022
Hm, I think calling |
Co-authored-by: Jake Hemstad <[email protected]>
|
@gpucibot merge |
rapids-bot bot
pushed a commit
that referenced
this pull request
Apr 24, 2022
This PR is for NVIDIA/spark-rapids#5029 and NVIDIA/spark-rapids#1870, which enables cuDF JNI to throw CUDA errors with specific error code. This PR relies on #10630, which exposes the CUDA error code and distinguishes fatal CUDA errors from the others. With this improvement, it is supposed to be easier to track CUDA errors triggered by JVM APIs. Authors: - Alfred Xu (https://github.com/sperlingxx) Approvers: - Jason Lowe (https://github.com/jlowe) URL: #10551
rapids-bot bot
pushed a commit
that referenced
this pull request
Jun 7, 2022
This PR is a follow-up PR of #10630, which is to improve the capture of fatal cuda errors in libcudf and cudf java package. 1. libcudf: Removes the redundent call of `cudaGetLastError` in throw_cuda_error, since the call returning the cuda error can be deemed as the first call. 2. JNI: Leverages similar logic to discern fatal cuda errors from catched exceptions. The check at the JNI level is necessary because fatal cuda errors due to rmm APIs can not be distinguished. 3. Add C++ unit test for the capture of fatal cuda error 4. Add Java unit test for the capture of fatal cuda error Authors: - Alfred Xu (https://github.com/sperlingxx) Approvers: - Jake Hemstad (https://github.com/jrhemstad) - Jason Lowe (https://github.com/jlowe) URL: #10884
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.
Closes #10553
Improves
cudf::cuda_errorin two aspects:cudf::cuda_errorand corresponding error_code() function that returns the error codecuda::cuda_errorassticky_cuda_errorandcudart_error.sticky_cuda_errorrefers to fatal error on device.