deepujain opened a new pull request, #59047: URL: https://github.com/apache/spark/pull/59047
### What changes were proposed in this pull request? Add a whole-stage code generation section to `docs/sql-performance-tuning.md` explaining `spark.sql.codegen.hugeMethodLimit`, its default, and the fallback when a generated method exceeds the limit. The configuration remains internal. Fixes [SPARK-59779](https://issues.apache.org/jira/browse/SPARK-59779). **JIRA assignee for credit:** deepujain ### Why are the changes needed? When generated query code grows too large for HotSpot to JIT compile, users need a way to understand and tune the fallback. The existing knob is absent from the tuning guide. This documents when trying `8000` can help, while making clear that turning off whole-stage code generation can also hurt performance and should be measured on the user's workload. ### Does this PR introduce _any_ user-facing change? No. Documentation only; the default and execution behavior are unchanged. ### How was this patch tested? Built the non-API documentation with Ruby 3.3 and `SKIP_ERRORDOC=1 SKIP_API=1 bundle exec jekyll build`, then checked the rendered section and configuration table. Verified the default, version, bytecode comparison, and subtree fallback against `SQLConf`, `CodeGenerator`, and `WholeStageCodegenExec`. `git diff --check` passed. No code or generated API/error documentation changed. ### Was this patch authored or co-authored using generative AI tooling? Generated-by: OpenAI Codex (GPT-6) -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
