Github user mridulm commented on a diff in the pull request:
https://github.com/apache/incubator-spark/pull/641#discussion_r10000899
--- Diff: core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala
---
@@ -373,25 +375,26 @@ class DAGScheduler(
} else {
def removeStage(stageId: Int) {
// data structures based on Stage
- stageIdToStage.get(stageId).foreach { s =>
- if (running.contains(s)) {
+ for (stage <- stageIdToStage.get(stageId)) {
+ if (running.contains(stage)) {
logDebug("Removing running stage %d".format(stageId))
- running -= s
+ running -= stage
+ }
+ stageToInfos -= stage
+ for (shuffleDep <- stage.shuffleDep) {
+ shuffleToMapStage.remove(shuffleDep.shuffleId)
}
- stageToInfos -= s
- shuffleToMapStage.keys.filter(shuffleToMapStage(_) ==
s).foreach(shuffleId =>
- shuffleToMapStage.remove(shuffleId))
- if (pendingTasks.contains(s) && !pendingTasks(s).isEmpty) {
+ if (pendingTasks.contains(stage) &&
!pendingTasks(stage).isEmpty) {
logDebug("Removing pending status for stage
%d".format(stageId))
}
- pendingTasks -= s
--- End diff --
That is surprising !
I would assume searching for an entry in a Set is atleast a few dozen
opcodes, if not more - even if already searched (and so bpt's are favourable);
would still be much more expensive than splitting an if for sure - I have not
looked into scala's impl : but assuming reasonable implementation of 'if', this
is definitely unexpected :-)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
