Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cuda 11.1 cudart SO name fix #43716

Merged
merged 3 commits into from Oct 7, 2020
Merged

Conversation

@nluehr
Copy link
Contributor

@nluehr nluehr commented Oct 2, 2020

In CUDA 11.1 the soname of libcudart remains versioned as 11.0 for better backward comparability. This PR updates the configure scripts accordingly.

This should also address issue #43689

Attn @sanjoy

@gbaned gbaned self-assigned this Oct 2, 2020
@gbaned gbaned added this to Assigned Reviewer in PR Queue via automation Oct 2, 2020
@gbaned gbaned requested a review from sanjoy Oct 2, 2020
Copy link
Member

@sanjoy sanjoy left a comment

LGTM with a minor nit.

third_party/gpus/cuda_configure.bzl Show resolved Hide resolved
PR Queue automation moved this from Assigned Reviewer to Reviewer Requested Changes Oct 5, 2020
sanjoy
sanjoy approved these changes Oct 6, 2020
PR Queue automation moved this from Reviewer Requested Changes to Approved by Reviewer Oct 6, 2020
@jacobaustin123
Copy link

@jacobaustin123 jacobaustin123 commented Oct 6, 2020

This is currently preventing TF and Jax builds on 11.1 (at least with Ampere cards). Making these changes to the bazel config fixes the issue.

@tensorflow-copybara tensorflow-copybara merged commit 68a6fe0 into tensorflow:master Oct 7, 2020
10 of 13 checks passed
10 of 13 checks passed
@tensorflow-jenkins
Intel® oneDNN -- Community CI Build oneDNN unit tests failed
Details
@kokoro-team
MacOS CPU Python3 Internal CI build failed
Details
@kokoro-team
Ubuntu CPU Internal CI build failed
Details
@kokoro-team
Android Demo App Internal CI build successful
Details
@kokoro-team
Linux GPU Internal CI build successful
Details
@kokoro-team
TFLite Android Demo App Internal CI build successful
Details
@kokoro-team
TFLite Makefile Internal CI build successful
Details
@kokoro-team
TFLite iOS Tests Internal CI build successful
Details
@kokoro-team
Ubuntu Sanity Internal CI build successful
Details
@kokoro-team
Windows Bazel Internal CI build successful
Details
@kokoro-team
Windows Bazel GPU Internal CI build successful
Details
@googlebot
cla/google All necessary CLAs are signed
import/copybara Change imported to the internal review system
Details
PR Queue automation moved this from Approved by Reviewer to Merged Oct 7, 2020
perfinion added a commit to perfinion/addons that referenced this issue Oct 27, 2020
libcudart soname in CUDA 11.x is versioned as 11.0 for backward compatability.
This change is taken from tensorflow/tensorflow#43716

find_cuda_config.py was taken directly from TF (and reformatted with
black). cusparse changed versioning in cuda11 so that part is needed.
The other changes pulled in from TF is mostly more lib/header search
paths and a few style changes.

Signed-off-by: Jason Zaman <jason@perfinion.com>
seanpmorgan added a commit to tensorflow/addons that referenced this issue Nov 22, 2020
* Update cuda config for CUDA11.1

libcudart soname in CUDA 11.x is versioned as 11.0 for backward compatability.
This change is taken from tensorflow/tensorflow#43716

find_cuda_config.py was taken directly from TF (and reformatted with
black). cusparse changed versioning in cuda11 so that part is needed.
The other changes pulled in from TF is mostly more lib/header search
paths and a few style changes.

Signed-off-by: Jason Zaman <jason@perfinion.com>

* * Fix flake8 and formatting errors

Co-authored-by: Sean Morgan <seanmorgan@outlook.com>
jrruijli pushed a commit to jrruijli/addons that referenced this issue Dec 23, 2020
* Update cuda config for CUDA11.1

libcudart soname in CUDA 11.x is versioned as 11.0 for backward compatability.
This change is taken from tensorflow/tensorflow#43716

find_cuda_config.py was taken directly from TF (and reformatted with
black). cusparse changed versioning in cuda11 so that part is needed.
The other changes pulled in from TF is mostly more lib/header search
paths and a few style changes.

Signed-off-by: Jason Zaman <jason@perfinion.com>

* * Fix flake8 and formatting errors

Co-authored-by: Sean Morgan <seanmorgan@outlook.com>
@nluehr nluehr deleted the cuda_11.1_fix branch Jan 13, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
PR Queue
  
Merged
Linked issues

Successfully merging this pull request may close these issues.

None yet

7 participants