Github user nongli commented on a diff in the pull request:
https://github.com/apache/spark/pull/11454#discussion_r54660492
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetReadBenchmark.scala
---
@@ -219,11 +216,43 @@ object ParquetReadBenchmark {
/*
Intel(R) Core(TM) i7-4870HQ CPU @ 2.50GHz
Int and String Scan: Best/Avg Time(ms) Rate(M/s)
Per Row(ns) Relative
-
-------------------------------------------------------------------------------
- SQL Parquet Reader 1381 / 1679 7.6
131.7 1.0X
- SQL Parquet MR 2005 / 2177 5.2
191.2 0.7X
- SQL Parquet Vectorized 919 / 1044 11.4
87.6 1.5X
- ParquetReader 1035 / 1163 10.1
98.7 1.3X
+
-------------------------------------------------------------------------------------------
+ SQL Parquet Vectorized 1025 / 1180 10.2
97.8 1.0X
+ SQL Parquet MR 2157 / 2222 4.9
205.7 0.5X
+ SQL Parquet Non-vectorized 1450 / 1466 7.2
138.3 0.7X
+ ParquetReader 1005 / 1022 10.4
95.9 1.0X
--- End diff --
The last one is the 1.6 parquet reader. What this is comparing, some what
randomly, is the 1.6 reader without spark (last row) vs the vectorized reader
with all of spark (first row).
---
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]
