Github user maropu commented on the issue:
https://github.com/apache/spark/pull/19021
Just for your info, again, I looked into this issue in TPC-DS quries; I
added [some
code](https://github.com/apache/spark/compare/master...maropu:SPARK-21603-FOLLOWUP-3)
to check the actual bytecode size of these quries and I found the gen'd
function in Q17/Q66 only had too-long bytecode over `8000`:
```
===== TPCDS QUERY BENCHMARK OUTPUT FOR q17 =====
17/08/23 14:45:02 WARN CodeGenerator:
GeneratedClass.agg_doAggregateWithKeys is too large to do JIT compilation on
HotSpot; the size of agg_doAggregateWithKeys is 17665; the limit is 8000
===== TPCDS QUERY BENCHMARK OUTPUT FOR q66 =====
17/08/23 14:55:39 WARN CodeGenerator:
GeneratedClass.agg_doAggregateWithKeys is too large to do JIT compilation on
HotSpot; the size of agg_doAggregateWithKeys is 11012; the limit is 8000
17/08/23 14:55:39 WARN CodeGenerator:
GeneratedClass.agg_doAggregateWithKeys is too large to do JIT compilation on
HotSpot; the size of agg_doAggregateWithKeys is 13420; the limit is 8000
17/08/23 14:55:39 WARN CodeGenerator:
GeneratedClass.agg_doAggregateWithKeys is too large to do JIT compilation on
HotSpot; the size of agg_doAggregateWithKeys is 16641; the limit is 8000
```
BTW, why we don't check if gen'd bytecode size is over `8000` directly
instead of code line num. in #18810? cc: @gatorsmile @viirya @kiszk
---
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]
