marin-ma commented on code in PR #12089:
URL: https://github.com/apache/gluten/pull/12089#discussion_r3240350057
##########
gluten-substrait/src/main/scala/org/apache/gluten/config/GlutenConfig.scala:
##########
@@ -1061,6 +1063,14 @@ object GlutenConfig extends ConfigRegistry {
.checkValue(v => v >= 0 && v <= 1, "Buffer reallocation threshold must
between [0, 1]")
.createWithDefault(0.25)
+ val COLUMNAR_SHUFFLE_EVICT_PARTITION_SIZE =
+ buildConf("spark.gluten.sql.columnar.shuffle.evictPartitionSize")
Review Comment:
I would suggest use
`spark.gluten.sql.columnar.shuffle.partitionBufferEvictThreshold` for this
configuration.
##########
gluten-substrait/src/main/scala/org/apache/gluten/config/GlutenConfig.scala:
##########
@@ -1061,6 +1063,14 @@ object GlutenConfig extends ConfigRegistry {
.checkValue(v => v >= 0 && v <= 1, "Buffer reallocation threshold must
between [0, 1]")
.createWithDefault(0.25)
+ val COLUMNAR_SHUFFLE_EVICT_PARTITION_SIZE =
+ buildConf("spark.gluten.sql.columnar.shuffle.evictPartitionSize")
+ .doc(
+ "For Velox hash shuffle writer, evict partition buffers larger than
this threshold " +
+ "after splitting an input batch.")
+ .intConf
+ .createWithDefault(256 * 1024)
Review Comment:
Can you set the default value to -1 to disable this eviction behaviour by
default?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
