Default CUDA host compiler the same as CXX compiler #3957
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
ye-luo
force-pushed
the
update-cudatoolkit
branch
from
64096fb
to
e890cc7
Compare
|
Test this please |
|
Test this please |
1) Require CUDA 11 2) -allow-unsupported-compiler by default.
ye-luo
force-pushed
the
update-cudatoolkit
branch
from
a8904de
to
e41530d
Compare
|
Test this please |
prckent
approved these changes
Apr 24, 2022
|
Test this please |
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Outdated suggestions cannot be applied.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.
Proposed changes
I noticed an issue with upstream clang which changes its default to PIE enabled.
Our nvcc setting default using gcc as the host compiler, this caused issue at linking on SLES 15 machines.
gcc doesn't have PIE enabled by default. This choice is made by the linux distro. Ubuntu has PIE enabled by default.
One fix is to pass -fPIE via CMAKE_CUDA_FLAGS but I dislike this solution.
Instead, I choose to set CMAKE_CXX_COMPILER as the default value of CMAKE_CUDA_HOST_COMPILER if user doesn't specify one.
This change needs additional changes.
What type(s) of changes does this code introduce?
Does this introduce a breaking change?
What systems has this change been tested on?
epyc-server/bora/jlse/perlmutter
Checklist