xuanyuanking commented on a change in pull request #26040: [SPARK-9853][Core]
Optimize shuffle fetch of continuous partition IDs
URL: https://github.com/apache/spark/pull/26040#discussion_r332982514
##########
File path:
common/network-shuffle/src/main/java/org/apache/spark/network/shuffle/protocol/FetchShuffleBlocks.java
##########
@@ -35,20 +35,27 @@
// The length of mapIds must equal to reduceIds.size(), for the i-th mapId
in mapIds,
// it corresponds to the i-th int[] in reduceIds, which contains all reduce
id for this map id.
public final long[] mapIds;
+ // When batchFetchEnabled=true, reduceIds[i] contains 2 elements:
startReduceId (inclusive) and
+ // endReduceId (exclusive) for the mapper mapIds[i].
+ // When batchFetchEnabled=false, reduceIds[i] contains all the reduce IDs
that mapper mapIds[i]
+ // needs to fetch.
public final int[][] reduceIds;
+ public final boolean batchFetchEnabled;
public FetchShuffleBlocks(
String appId,
String execId,
int shuffleId,
long[] mapIds,
- int[][] reduceIds) {
+ int[][] reduceIds,
+ boolean batchFetchEnabled) {
this.appId = appId;
this.execId = execId;
this.shuffleId = shuffleId;
this.mapIds = mapIds;
this.reduceIds = reduceIds;
assert(mapIds.length == reduceIds.length);
+ this.batchFetchEnabled = batchFetchEnabled;
Review comment:
Thanks, done in 3b6c5a6.
----------------------------------------------------------------
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]
