Github user Ngone51 commented on a diff in the pull request:
https://github.com/apache/spark/pull/19285#discussion_r162962203
--- Diff:
core/src/main/scala/org/apache/spark/storage/memory/MemoryStore.scala ---
@@ -233,17 +235,13 @@ private[spark] class MemoryStore(
}
if (keepUnrolling) {
- // We successfully unrolled the entirety of this block
- val arrayValues = vector.toArray
- vector = null
- val entry =
- new DeserializedMemoryEntry[T](arrayValues,
SizeEstimator.estimate(arrayValues), classTag)
- val size = entry.size
+ // We need more precise value
+ val size = valuesHolder.esitimatedSize(false)
--- End diff --
`roughly = false` tells more than estimating the size of vector, which is
'unroll has finished'. So, `storeValue` will not be called anymore.
And for `heavy work`(under my understanding, mostly due to
`SizeEstimator.estimate(arrayValues)`), yeah, I agree with you. Those `heavy
work` would be done whether we create an entry or not or whether the entry put
into the memory store or not finally.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
