srowen commented on a change in pull request #26469: [SPARK-29844] Improper
unpersist strategy in ml.recommendation.ASL.train
URL: https://github.com/apache/spark/pull/26469#discussion_r344747970
##########
File path: mllib/src/main/scala/org/apache/spark/ml/recommendation/ALS.scala
##########
@@ -1041,13 +1041,13 @@ object ALS extends DefaultParamsReadable[ALS] with
Logging {
.persist(finalRDDStorageLevel)
if (finalRDDStorageLevel != StorageLevel.NONE) {
userIdAndFactors.count()
- itemIdAndFactors.count()
- itemFactors.unpersist()
userInBlocks.unpersist()
userOutBlocks.unpersist()
- itemInBlocks.unpersist()
itemOutBlocks.unpersist()
blockRatings.unpersist()
+ itemIdAndFactors.count()
Review comment:
Yes, but itemFactors depends on the previous user RDDs, including
userOutBlocks. I think this is done on purpose. EDIT: Hm OK but it's persisted,
so shouldn't matter as long as itemFactors is unpersisted afterwards. This is a
bit of an optimization right? just tries to free up more space before
persisting the item stuff.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
