-
-
Notifications
You must be signed in to change notification settings - Fork 427
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
Update cuda arch and change cuda version #5473
Update cuda arch and change cuda version #5473
Conversation
Looks like it only solves the |
eb67577
to
2d6294e
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks!
retest lsu |
3d4627a
to
3e081ef
Compare
retest lsu |
3e081ef
to
90c298c
Compare
90c298c
to
2a0cf4c
Compare
eeabcb3
to
88acaf8
Compare
04776a9
to
53d0285
Compare
retest lsu |
Thanks @aurianer for updating this! It looks good except that there were odd failures on the release build with nvcc (executables not found even though they were built). My guess is that it was unfortunate timing with the maintenance on rostam, but let's see. I've retriggered the builds on rostam. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks!
Still have to investigate a compilation error which triggers the CI error (executables do not exist) so not ready to merge |
53d0285
to
b3cd2bf
Compare
Removing the variadic ... of the tag_override_dispatch makes it compile.
The version change should be done manually in a PR to address all the problems and not by default when the default module on rostam is upgraded.
ae487e5
to
cbbc318
Compare
Mismatch of pointers between __global__ and __host__ __device__.
cbbc318
to
d5c778b
Compare
retest lsu |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good to me, and works nicely on rostam!
I'd just like to wait for the CUDA configuration on CSCS to verify that it doesn't break with older CUDA versions. Piz Daint maintenance is almost over and should be available again later today, but it may first require changes to the CUDA configuration on Piz Daint.
retest cscs |
- Downgrade the cuda version on rostam to avoid the errors currently on masterhttps://cdash.cscs.ch/viewBuildError.php?buildid=172746
Created this ticket #5472 to fix the problem with cuda/11.3EDIT: