gagafunctor commented on a change in pull request #23983: [SPARK-26881][mllib]
Heuristic for tree aggregate depth
URL: https://github.com/apache/spark/pull/23983#discussion_r264168029
##########
File path:
mllib/src/test/scala/org/apache/spark/mllib/linalg/distributed/RowMatrixSuite.scala
##########
@@ -101,6 +101,17 @@ class RowMatrixSuite extends SparkFunSuite with
MLlibTestSparkContext {
}
}
+ test("getTreeAggregateIdealDepth") {
+ val nbPartitions = 10000
+ val vectors = sc.emptyRDD[Vector]
+ .repartition(nbPartitions)
+ vectors.conf.set("spark.driver.maxResultSize", "10g")
Review comment:
I checked the maxResultSize parameter value in the test and it's the default
one (1gig, e.g 1024 Mb) -> I'll change the numPartitions and object size to:
- numPartitions -> 100
- objectSize -> 100 to get a desired depth of 2 (because sqrt(100) * 100 <=
1024)
- objectSize -> 110 to get a desired depth of 3 (because sqrt(100) * 110 >
1024 and math.pow(100, 0.33) * 110 > 1024)
So the (semantically) same test doesn't require a change in the
sharedContext spark conf
----------------------------------------------------------------
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]
