Github user mridulm commented on a diff in the pull request:
https://github.com/apache/spark/pull/16989#discussion_r116830282
--- Diff: core/src/main/scala/org/apache/spark/scheduler/MapStatus.scala ---
@@ -193,8 +216,21 @@ private[spark] object HighlyCompressedMapStatus {
} else {
0
}
+ val threshold =
+ SparkEnv.get.conf.getLong("spark.shuffle.accurateBlkThreshold", 100
* 1024 * 1024) max
+
SparkEnv.get.conf.getInt("spark.shuffle.accurateBlkThresholdByTimesAvg", 2) *
avgSize
--- End diff --
Agree about moving it to `org.apache.spark.internal.config`.
I am fine with it either way - a constant size does have the potential for
degenerating into `CompressedMapStatus` for large shuffle's.
Btw, can we explicitly use math.max() ? instead of `v1 max v2` ? I thought
that was the idiom used in spark ?
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
