Skip to content

Commit

Permalink
Move cluster cols to config file
Browse files Browse the repository at this point in the history
Signed-off-by: Partho Sarthi <[email protected]>
  • Loading branch information
parthosa committed Jul 26, 2024
1 parent 52e1383 commit aa1149d
Show file tree
Hide file tree
Showing 2 changed files with 12 additions and 3 deletions.
5 changes: 2 additions & 3 deletions user_tools/src/spark_rapids_pytools/rapids/qualification.py
Original file line number Diff line number Diff line change
Expand Up @@ -787,9 +787,8 @@ def _infer_clusters_for_apps(self, cluster_info_df: pd.DataFrame) -> None:
if self.ctxt.get_ctxt('cpuClusterProxy') is not None or not self.ctxt.platform.cluster_inference_supported:
self.logger.info('CPU cluster is already set. Skipping cluster inference.')
return
cpu_cluster_cols = ['Num Executor Nodes', 'Executor Instance', 'Cores Per Executor']
gpu_cluster_cols = ['Recommended Num Executor Nodes', 'Recommended Executor Instance',
'Recommended Cores Per Executor']
cpu_cluster_cols = self.ctxt.get_value('local', 'output', 'clusterInference', 'cpuClusterColumns')
gpu_cluster_cols = self.ctxt.get_value('local', 'output', 'clusterInference', 'gpuClusterColumns')
# == Infer CPU clusters per app ==
# Drop GPU/Recommended columns to infer the CPU cluster information
cpu_cluster_df = cluster_info_df.drop(columns=gpu_cluster_cols, errors='ignore')
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -341,6 +341,16 @@ local:
dstCol: 'Estimated GPU Speedup'
- srcCol: 'appDuration_pred'
dstCol: 'Estimated GPU Duration'
clusterInference:
cpuClusterColumns:
- 'Num Executor Nodes'
- 'Executor Instance'
- 'Cores Per Executor'
gpuClusterColumns:
- 'Recommended Num Executor Nodes'
- 'Recommended Executor Instance'
- 'Recommended Cores Per Executor'

platform:
shortName: 'qual'
outputDir: qual-tool-output
Expand Down

0 comments on commit aa1149d

Please sign in to comment.