Github user viirya commented on the pull request:
https://github.com/apache/spark/pull/12729#issuecomment-216410360
Benchmark on dataset.
codes:
test("typed aggregation: TypedAggregator, expr, expr") {
val N = 50L << 20
runBenchmark("typed aggregation", N) {
sqlContext.range(N).map(x => (x & 65535, x)).groupByKey(_._1).agg(
typed.sumLong(_._2),
typed.sumLong(_._2),
typed.sumLong(_._2),
typed.sumLong(_._2)).collect()
}
old:
Java HotSpot(TM) 64-Bit Server VM 1.8.0_25-b17 on Linux
3.13.0-57-generic
Westmere E56xx/L56xx/X56xx (Nehalem-C)
typed aggregation: Best/Avg Time(ms) Rate(M/s)
Per Row(ns) Relative
-------------------------------------------------------------------------------------------
typed aggregation codegen=false 37749 / 39395 1.4
720.0 1.0X
typed aggregation codegen=true 29611 / 30014 1.8
564.8 1.3X
new:
Java HotSpot(TM) 64-Bit Server VM 1.8.0_25-b17 on Linux
3.13.0-57-generic
Westmere E56xx/L56xx/X56xx (Nehalem-C)
typed aggregation: Best/Avg Time(ms) Rate(M/s)
Per Row(ns) Relative
-------------------------------------------------------------------------------------------
typed aggregation codegen=false 38211 / 39191 1.4
728.8 1.0X
typed aggregation codegen=true 27384 / 27595 1.9
522.3 1.4X
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
