wbo4958 commented on code in PR #44690:
URL: https://github.com/apache/spark/pull/44690#discussion_r1473808064
##########
core/src/main/scala/org/apache/spark/resource/ResourceAllocator.scala:
##########
@@ -20,6 +20,49 @@ package org.apache.spark.resource
import scala.collection.mutable
import org.apache.spark.SparkException
+import org.apache.spark.resource.ResourceAmountUtils.ONE_ENTIRE_RESOURCE
+
+private[spark] object ResourceAmountUtils {
+ /**
+ * Using "double" to do the resource calculation may encounter a problem of
precision loss. Eg
+ *
+ * scala> val taskAmount = 1.0 / 9
+ * taskAmount: Double = 0.1111111111111111
+ *
+ * scala> var total = 1.0
+ * total: Double = 1.0
+ *
+ * scala> for (i <- 1 to 9 ) {
+ * | if (total >= taskAmount) {
+ * | total -= taskAmount
+ * | println(s"assign $taskAmount for task $i, total left: $total")
+ * | } else {
+ * | println(s"ERROR Can't assign $taskAmount for task $i, total
left: $total")
+ * | }
+ * | }
+ * assign 0.1111111111111111 for task 1, total left: 0.8888888888888888
+ * assign 0.1111111111111111 for task 2, total left: 0.7777777777777777
+ * assign 0.1111111111111111 for task 3, total left: 0.6666666666666665
+ * assign 0.1111111111111111 for task 4, total left: 0.5555555555555554
+ * assign 0.1111111111111111 for task 5, total left: 0.44444444444444425
+ * assign 0.1111111111111111 for task 6, total left: 0.33333333333333315
+ * assign 0.1111111111111111 for task 7, total left: 0.22222222222222204
+ * assign 0.1111111111111111 for task 8, total left: 0.11111111111111094
+ * ERROR Can't assign 0.1111111111111111 for task 9, total left:
0.11111111111111094
+ *
+ * So we multiply ONE_ENTIRE_RESOURCE to convert the double to long to avoid
this limitation.
+ * Double can display up to 16 decimal places, so we set the factor to
+ * 10, 000, 000, 000, 000, 000L.
+ */
+ final val ONE_ENTIRE_RESOURCE: Long = 10000000000000000L
Review Comment:
Hi @srowen,
Thank you for your review. You've suggested a straightforward solution to
address the issue of double precision by actively sacrificing some accuracy,
and it's certainly doable. However, this raises a question about the precision
that Spark should define. For instance, if a user sets the value to
0.1111111111111, which value should Spark use for the calculation: 0.111,
0.1111, or something else? If we simply select 4 decimal places, what happens
if a user wants to set the value to 0.000001? In that case, the value would be
converted to 0 and might raise an exception. But It seems that a configuration
option could resolve this problem.
My initial thought was that the resource fraction calculation should work
regardless of the value set by the user, and the proposed approach in the pull
request to convert the double to a Long can achieve this goal.
Nevertheless, if you and other committers believe that a change is
necessary, I am open to submitting a pull request for the master branch.
--
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]
