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

Remove total core second heuristic and filter apps only in top candidate view #1376

Merged
merged 5 commits into from
Oct 9, 2024

Conversation

cindyyuanjiang
Copy link
Collaborator

@cindyyuanjiang cindyyuanjiang commented Oct 5, 2024

Fixes #1367

Changes

  • Remove total core seconds filter from the qualification_summary.csv file.
  • Apply total core second threshold filter only in Stdout Top Candidate View table when --filter_apps=TOP_CANDIDATES. Nothing should be affected when --filter_apps=ALL.

Testing

Cmd: spark_rapids qualification -v -e <my-event-logs> --filter_apps 'ALL'
Stdout:

+----+-----------------------------------------------+--------------------------------+-----------------+------------------------------------+-------------------------------------+------------------------------------+
|    | App Name                                      | App ID                         | Estimated GPU   | Qualified                          | Full Cluster                        | GPU Config                         |
|    |                                               |                                | Speedup         | Cluster                            | Config                              | Recommendation                     |
|    |                                               |                                | Category**      | Recommendation                     | Recommendations*                    | Breakdown*                         |
|----+-----------------------------------------------+--------------------------------+-----------------+------------------------------------+-------------------------------------+------------------------------------|
|  2 | sanity_test                                   | application_1673503196578_0001 | Small           | 4 x Node with 8 vCPUs (1 L4 each)  | application_1673503196578_0001.conf | application_1673503196578_0001.log |
|  3 | sanity_test                                   | application_1673503196578_0001 | Small           | 4 x Node with 8 vCPUs (1 L4 each)  | application_1673503196578_0001.conf | application_1673503196578_0001.log |
|  8 | Spark shell                                   | local-1713904114464            | Not Recommended | 1 x Node with 64 vCPUs (1 L4 each) | local-1713904114464.conf            | local-1713904114464.log            |
|  7 | Spark shell                                   | local-1713464523119            | Not Recommended | 4 x Node with 8 vCPUs (1 L4 each)  | Does not exist, see log for errors  | Does not exist, see log for errors |
+----+-----------------------------------------------+--------------------------------+-----------------+------------------------------------+-------------------------------------+------------------------------------+
* Config Recommendations can be found in <path>.
** Estimated GPU Speedup Category assumes the user is using the node type recommended and config recommendations with the same size cluster as was used with the CPU side.

Report Summary:
----------------------  --
Total applications      6
Processed applications  4
Top candidates           0
----------------------  --

Cmd: spark_rapids qualification -v -e <my-event-logs>
Stdout:

Qualification tool found no qualified applications after applying the filters.
See the CSV file for the full report or disable the filters.

Report Summary:
----------------------  --
Total applications      6
Processed applications  4
Top candidates           0
----------------------  --

Signed-off-by: cindyyuanjiang <cindyj@nvidia.com>
Signed-off-by: cindyyuanjiang <cindyj@nvidia.com>
Signed-off-by: cindyyuanjiang <cindyj@nvidia.com>
@cindyyuanjiang cindyyuanjiang self-assigned this Oct 5, 2024
@cindyyuanjiang cindyyuanjiang added bug Something isn't working user_tools Scope the wrapper module running CSP, QualX, and reports (python) usability track issues related to the Tools's user experience labels Oct 5, 2024
@tgravescs
Copy link
Collaborator

tgravescs commented Oct 7, 2024

can you remove bullet 1 in the description. or reword it to say from the application_summary file, it was a bit confusing when I read it until I read the second bullet because I thought this some how changed to remove it fully. most people won't know when you say heuristic it doesn't apply to everything

tgravescs
tgravescs previously approved these changes Oct 7, 2024
@cindyyuanjiang
Copy link
Collaborator Author

can you remove bullet 1 in the description. or reword it to say from the application_summary file, it was a bit confusing when I read it until I read the second bullet because I thought this some how changed to remove it fully. most people won't know when you say heuristic it doesn't apply to everything

Thanks @tgravescs! I updated the PR description.

Copy link
Collaborator

@parthosa parthosa left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @cindyyuanjiang. Added a minor comment.

Also after this change in, I was thinking that we should improve usage of Heuristics + Filtering and what should be used when. Created a tracking issue #1379

user_tools/src/spark_rapids_tools/tools/top_candidates.py Outdated Show resolved Hide resolved
Signed-off-by: cindyyuanjiang <cindyj@nvidia.com>
Signed-off-by: cindyyuanjiang <cindyj@nvidia.com>
Copy link
Collaborator

@parthosa parthosa left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @cindyyuanjiang. LGTME.

@cindyyuanjiang cindyyuanjiang merged commit eabb6c1 into NVIDIA:dev Oct 9, 2024
14 checks passed
@cindyyuanjiang cindyyuanjiang deleted the spark-rapids-tools-1367 branch October 9, 2024 23:27
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working usability track issues related to the Tools's user experience user_tools Scope the wrapper module running CSP, QualX, and reports (python)
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[BUG] Change qualification for core seconds heuristic to not disqualify
3 participants