January 23, 2026: Weekly Status Update in Gluten

[GlutenPerfBot]

This weekly update is generated by LLMs. You're welcome to join our Github for in-depth discussions.

Overall Activity Summary

The Apache Gluten project has been highly active over the past 7 days with 47 pull requests and 24 issues, showing strong momentum across multiple fronts. Key themes include Spark 4.1 compatibility improvements, Delta Lake optimizations, ANSI mode support, and infrastructure enhancements. The community is actively addressing performance issues, compatibility challenges, and expanding platform support.

Key Ongoing Projects

Spark 4.1 Compatibility & ANSI Mode Support

• Major effort led by @baibaichen to track and fix Spark 4.1.x unit test failures ([VL] Track on Spark-4.1.x failed unit tests #11400), with significant progress on union partitioning support and Python 3.10 migration
• @malinjawi is driving ANSI mode support implementation ([VL] Add ANSI mode support #10134) with string-to-boolean casting completed ([GLUTEN-10134][VL] Add ANSI mode support for cast string to boolean #11437)
• @ReemaAlzaid updated CI to Python 3.10 for Spark 4.1 compatibility ([VL] Update CI Python to 3.10 for Spark 4.1 and enable ArrowEvalPythonExecSuite tests #11481)

Delta Lake Performance Optimizations

• @zhztheplayer is leading comprehensive Delta write optimizations ([VL] Delta Lake write #10215) with multiple PRs addressing native statistics tracking ([GLUTEN-10215][VL] Delta write: Native statistics tracker to avoid C2R overhead #11419), redundant C2R2C transitions ([GLUTEN-10215][VL] Delta Write: Fix redundant C2R2C transition #11478), and V1 fallback writers ([GLUTEN-10215][VL] Delta Write: Offload V1 fallback writers #11479)
• @FelixYBW identified and is addressing OOM issues during Delta table creation (OOM due to Hash based shuffle fallback #11421)

Infrastructure & Build Improvements

• @zhouyuan added RHEL 9.7 support ([VL] Adding RHEL 9.7 support #11460) and Maven version upgrades ([GLUTEN-11432][VL] Bump Maven from 3.9.2 to 3.9.12 #11431, [GLUTEN-11432][FOLLOWUP] Bump project level Maven to 3.9.12 #11470)
• @jinchengchenghh simplified cuDF build configuration (build(cudf): Simplify cuDF build configuration #11407) and fixed CI reporting issues ([VL][CI] Fix CI reporting success despite command failures by adding 'set -e' #11462)
• Daily Velox version updates are being maintained by @GlutenPerfBot

Priority Items

Critical Performance Issues

• TPCDS q72 uses gluten calculation much worse than spark #8417: TPC-DS Q72 performance regression needs immediate attention - Gluten showing significantly worse performance than vanilla Spark
• OOM due to Hash based shuffle fallback #11421: OOM during Delta table creation with clustering - affects production workloads
• Hive partitioned output fails as "exit code 134" #11397: Hive partitioned output failures with exit code 134

Compatibility & Stability

• [VL] Flaky test UnsafeColumnarBuildSideRelationTest in CI #11406: Flaky test in UnsafeColumnarBuildSideRelationTest causing CI instability
• [VL] Jvm crashed on JavaRssClient release #11473: JVM crashes in JavaRssClient release - production stability issue
• [CH] UT failures from ClickHouseAdaptiveQueryExecSuite #11369: ClickHouse adaptive query execution test failures

Build & Platform Support

• [BUG][VL] Inconsistent C++ Compiler Flags for aarch64 Causing xsimd initialization errors #11390: ARM64 compiler flag inconsistencies causing xsimd initialization errors
• [VL][BUG] task crash when use Dynamically sizing off-heap memory #11445: Dynamic off-heap sizing configuration issues

Notable Discussions

Performance Benchmarking Discussion

• [Question] TPC-DS 1TB Benchmarking results for Non-Partitioned Delta tables with Velox Backend #11463: Community member @shadowmmu seeking TPC-DS 1TB benchmarking results for non-partitioned Delta tables with Velox backend, highlighting need for more comprehensive performance data on lakehouse formats

Brand Compliance Initiative

• Track & remediate third‑party branding references #11438: @weiting-chen leading effort to ensure proper "Apache Gluten" branding across vendor documentation, with several vendors already updated

Emerging Trends

1. Lakehouse Format Maturation: Significant focus on Delta Lake optimizations with native write path improvements and performance tuning
2. Spark 4.1 Migration: Major push towards full Spark 4.1 compatibility with ANSI mode support becoming critical
3. Platform Diversification: Expanding support for ARM64, RHEL 9.7, and different JDK versions
4. Memory Management: Multiple issues around off-heap memory management and OOM prevention
5. CI/CD Stabilization: Ongoing efforts to reduce flaky tests and improve build reliability

Good First Issues

#11400: Spark 4.1 Unit Test Fixes

• What: Help fix failing Spark 4.1 unit tests across various components
• Skills Needed: Scala, Spark internals, test debugging
• Why Good: Well-documented with specific test cases, clear acceptance criteria, and community support

#10134: ANSI Mode Support Implementation

• What: Implement ANSI-compliant casting functions for string-to-timestamp, string-to-date conversions
• Skills Needed: C++, Velox functions, Spark SQL semantics
• Why Good: Clear specifications with Spark reference implementations, good for learning Velox function development

#11417: Qualification Tool Enhancement

• What: Add lakehouse format detection (Iceberg, Delta, Hudi, Paimon) to the qualification tool
• Skills Needed: Java, pattern matching, data format knowledge
• Why Good: Self-contained tool enhancement with clear requirements and test coverage expectations

#8984: Varchar to Timestamp Casting Corner Cases

• What: Fix edge cases in varchar to timestamp casting uncovered by Spark tests
• Skills Needed: C++, date/time parsing, edge case analysis
• Why Good: Focused scope with existing test cases, good for understanding Velox type system