Set USE_KINETO=1#49897
Conversation
Summary: Resend of #49201 Test Plan: see 49201 [ghstack-poisoned]
💊 CI failures summary and remediationsAs of commit 4aeceac (more details on the Dr. CI page):
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions to the (internal) Dr. CI Users group. |
|
I reverted earlier 49201 because of some ddp test failures in one of the master builds, will investigate |
|
@ilia-cher please make sure to check that your PR does not add any direct dependencies against libcuda.so (i.e. PyTorch build with CUDA should be loadable on systems that do not have CUDA driver installed) |
yes, I think we should be fine - we only compile cupti and kineto statically, and CI has non-CUDA builds as well |
|
CI looks good (mod. two unrelated failures) |
Summary: Resend of #49201 Test Plan: see 49201 Differential Revision: [D25717102](https://our.internmc.facebook.com/intern/diff/D25717102) [ghstack-poisoned]
Summary: Resend of #49201 Test Plan: see 49201 Differential Revision: [D25717102](https://our.internmc.facebook.com/intern/diff/D25717102) [ghstack-poisoned]
Summary: Resend of #49201 Test Plan: see 49201 Differential Revision: [D25717102](https://our.internmc.facebook.com/intern/diff/D25717102) [ghstack-poisoned]
Summary: Resend of #49201 Test Plan: see 49201 Differential Revision: [D25717102](https://our.internmc.facebook.com/intern/diff/D25717102) [ghstack-poisoned]
Summary: Resend of #49201 Test Plan: see 49201 Differential Revision: [D25717102](https://our.internmc.facebook.com/intern/diff/D25717102) [ghstack-poisoned]
Codecov Report
@@ Coverage Diff @@
## gh/ilia-cher/94/base #49897 +/- ##
========================================================
- Coverage 81.02% 81.02% -0.01%
========================================================
Files 1916 1916
Lines 209349 209349
========================================================
- Hits 169628 169621 -7
- Misses 39721 39728 +7 |
|
@ilia-cher merged this pull request in e34992e. |
| message(STATUS " KINETO_LIBRARY_TYPE = ${KINETO_LIBRARY_TYPE}") | ||
| message(STATUS " CUDA_SOURCE_DIR = ${CUDA_SOURCE_DIR}") | ||
|
|
||
| if(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti_static.a) |
There was a problem hiding this comment.
Is it possible to use CUDA_cupti_LIBRARY found by FindCUDA here? Also, see #51026, USE_CUPTI_SO should be given preference.
Summary: Pull Request resolved: pytorch#49897 Resend of pytorch#49201 Test Plan: see 49201 Reviewed By: malfet Differential Revision: D25717102 Pulled By: ilia-cher fbshipit-source-id: 5e794a7f5fe160ca64ac9d190c4fd3e8f1e443e6
Stack from ghstack:
Summary:
Resend of #49201
Test Plan:
see 49201
Differential Revision: D25717102