Add Cost Estimator Using Past Statistics for Schedule Generator #3156
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Yicong-Huang
approved these changes
Dec 20, 2024
Comment on lines
+89
to
+95
| SqlServer | ||
| .getInstance( | ||
| StorageConfig.jdbcUrl, | ||
| StorageConfig.jdbcUsername, | ||
| StorageConfig.jdbcPassword | ||
| ) | ||
| .createDSLContext() |
There was a problem hiding this comment.
@bobbai00 is this the correct way to use connection pool?
Comment on lines
+128
to
+145
| val eid = latestSuccessfulExecution.get.eId | ||
| val rawStats = context | ||
| .select( | ||
| WORKFLOW_RUNTIME_STATISTICS.OPERATOR_ID, | ||
| WORKFLOW_RUNTIME_STATISTICS.TIME, | ||
| WORKFLOW_RUNTIME_STATISTICS.DATA_PROCESSING_TIME, | ||
| WORKFLOW_RUNTIME_STATISTICS.CONTROL_PROCESSING_TIME | ||
| ) | ||
| .from(WORKFLOW_RUNTIME_STATISTICS) | ||
| .where( | ||
| WORKFLOW_RUNTIME_STATISTICS.WORKFLOW_ID | ||
| .eq(widAsUInteger) | ||
| .and(WORKFLOW_RUNTIME_STATISTICS.EXECUTION_ID.eq(eid)) | ||
| ) | ||
| .orderBy(WORKFLOW_RUNTIME_STATISTICS.TIME, WORKFLOW_RUNTIME_STATISTICS.OPERATOR_ID) | ||
| .fetchInto(classOf[WorkflowRuntimeStatistics]) | ||
| .asScala | ||
| .toList |
There was a problem hiding this comment.
if possible, can we use one query to get the latest successful execution's stats? we can avoid the transaction.
Comment on lines
+146
to
+158
| if (rawStats.isEmpty) { | ||
| None | ||
| } else { | ||
| val cumulatedStats = rawStats.foldLeft(Map.empty[String, Double]) { (acc, stat) => | ||
| val opTotalExecutionTime = acc.getOrElse(stat.getOperatorId, 0.0) | ||
| acc + (stat.getOperatorId -> (opTotalExecutionTime + (stat.getDataProcessingTime | ||
| .doubleValue() + stat.getControlProcessingTime.doubleValue()) / 1e9)) | ||
| } | ||
| Some(cumulatedStats) | ||
| } | ||
| } else { | ||
| None | ||
| } |
There was a problem hiding this comment.
try to use early return or other techniques (e.g., pattern matching) to reduce branches.
Comment on lines
+191
to
+195
| val region = Region( | ||
| id = RegionIdentity(0), | ||
| physicalOps = workflow.physicalPlan.operators, | ||
| physicalLinks = workflow.physicalPlan.links | ||
| ) |
There was a problem hiding this comment.
can we add a test case where the region only contains a sub-dag of the workflow? just to make sure it can extract region costs correctly.
Comment on lines
+140
to
+144
| val region = Region( | ||
| id = RegionIdentity(0), | ||
| physicalOps = workflow.physicalPlan.operators, | ||
| physicalLinks = workflow.physicalPlan.links | ||
| ) |
There was a problem hiding this comment.
can we add a test case where the region only contains a sub-dag of the workflow? just to make sure it can extract region costs correctly.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR introduces the
CostEstimatortrait which estimates the cost of a region, given some resource units.CostBasedScheduleGeneratorto calculate the cost of a schedule during search.A
DefaultCostEstimatorimplementation is also added, which uses past execution statistics to estimate the wall-clock runtime of a region: