-
Notifications
You must be signed in to change notification settings - Fork 35
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
Remove arguments related to cost-savings #1230
Remove arguments related to cost-savings #1230
Conversation
Signed-off-by: Ahmed Hussein (amahussein) <a@ahussein.me> Fixes NVIDIA#1229 - remove the legacy `spark_rapids_user_tools` cmd - remove qualification arguments related to cost-savings
Signed-off-by: Ahmed Hussein (amahussein) <a@ahussein.me> Fixes NVIDIA#1099 - disable grouping of results by row_name - the file `qualification_summary_full.csv` is omitted
Signed-off-by: Ahmed Hussein (amahussein) <a@ahussein.me>
Signed-off-by: Ahmed Hussein (amahussein) <a@ahussein.me>
* Fix node recommendation when CPU cluster cannot be determined Signed-off-by: Partho Sarthi <psarthi@nvidia.com> * Move cluster cols to config file Signed-off-by: Partho Sarthi <psarthi@nvidia.com> --------- Signed-off-by: Partho Sarthi <psarthi@nvidia.com>
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.
Thanks @amahussein for fixing these.
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.
Thanks @amahussein!
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.
Thanks @amahussein !
Thanks @parthosa for appending to this PR. |
Signed-off-by: Ahmed Hussein (amahussein) a@ahussein.me
Fixes #1229, Fixes #1099
This PR is to avoid errors that could be triggered by passing cost-savings arguments.
More cleaning of dead-code can be part of the parent issue- #1221
spark_rapids_user_tools
cmdqualification_summary_full.csv
is omittedThe following arguments in rapids_tools qualification cmd:
Fix Cluster Recommendation when CPU cluster cannot be created
Additionally, this PR fixes the issue where we do not generate a cluster recommendation when CPU cluster cannot be created (e.g., no matching executor instance found for the required number of cores).
Approach
Scala tool now generates a recommended GPU cluster per app basis NVIDIA/spark-rapids-tools#1188. For the case when CPU cluster is not provided, we should use the values from Scala tool output for our GPU cluster recommendation instead of python's cpu<->gpu core matching.
Output
Case 1: CPU cluster is not passed and we infer CPU cluster for each app
Logs (for each app):
Final Result:
Case 2: CPU cluster is passed as input (
--cluster <cluster>
)Logs (for all apps):
Final Result:
PR for this change: amahussein#13