-
Notifications
You must be signed in to change notification settings - Fork 282
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
add custom easyblock for AOMP (AMD OpenMP compiler) #2435
Conversation
This PR adds an AOMP EasyBlock to better pick `CUDA` support.
The AOMP build creates a symlink setup so that multiple installations can exist side-by-side. This is not needed by EasyBuild and is even making it difficult to remove and reinstall. To fix this, this commit adds a post install step which makes it the way EB expects.
Use only the '--cuda-compute-capabilties' build_option.
@akesandgren Would you be able to review this again? |
easybuild/easyblocks/a/aomp.py
Outdated
@@ -71,7 +70,6 @@ def configure_step(self): | |||
# Use the commandline / easybuild config option if given, else use | |||
# the value from the EC (as a default) | |||
cuda_cc = build_option('cuda_compute_capabilities') | |||
cuda_cc = cuda_cc or self.cfg['cuda_compute_capabilities'] |
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.
You should still allow for cuda_compute_capabilities to be specified in the easyconfig, i.e. this line should not be removed.
See tensorflow.py line 551 for a example on how to use it.
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.
I thought you wanted me to remove the EB capability of defining cuda_compute_capabilities
and that it had to be defined in the extra_vars
option.
I have added back in the capability to either use the command line parameter or variable defined in EB (copied from TF EB).
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.
No, it was just that you don't need to specify it in extra_vars since it is made available there from the framework.
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
Going in, thanks @nordmoen! |
This PR adds an AOMP EasyBlock to better pick
CUDA
support.