xylaaaaa commented on issue #69884: URL: https://github.com/apache/doris/issues/69884#issuecomment-6569729100
## TPC-DS rerun after basic statistics — 2026-11-02 **All 103 query files passed in one official regression suite. Delta's summed-median gap to Hive narrows from 26.1% to 16.5%; performance acceptance remains open.** | Sum of per-query medians (three alternating measured pairs) | Hive | Delta | Extra Delta time | | --- | ---: | ---: | ---: | | Before statistics | 232.545 s | 293.214 s | +26.1% | | After statistics | 225.408 s | 262.556 s | +16.5% | | After, 91 nonempty-result queries | 212.654 s | 246.057 s | +15.7% | The gap shrinks by **9.6 percentage points**. Delta's total falls 10.46%, while Hive's falls 3.07%. These are paired local ASAN/SF1 observations, not RELEASE/TPC-certified results or an isolated attribution of every change to statistics. The rerun reuses exactly the retained Parquet objects, the same 103 SQL files, four pipeline tasks, disabled SQL/file caches, 8-GiB/query, 120-second timeout and UTC. Q85 keeps the existing common fixed join order; Q30 keeps the same generator-column adaptation. All 24 Delta EXPLAIN cardinalities matched the independent converter manifest before timing; Hive's existing estimates were recorded without changing them. All **824 executions** were compared through the existing typed result comparator. TPC-H was not rerun. Observed improvements and remaining gaps: - **Q43:** Delta median 6.006 → 2.692 s; current Hive 2.442 s. Complete profiles show its hash build changes from 2,750,370 rows to 366, matching Hive. - **Q72:** Delta median 8.013 → 4.227 s; current Hive 4.000 s. Delta no longer builds from all 11,745,000 inventory rows; inventory scan output is now 1,460,940, matching Hive. - **Q22:** current Hive/Delta 8.367/11.684 s, still about 40% slower. Scan output/time and build sides match closely; aggregation/adaptive preaggregation and pipeline behavior still need isolation. - **Regressions remain visible:** Q4/Q11/Q74 lost earlier Delta advantages. A separate unprofiled Q4 rerun confirms 6.167/8.893 s (+44.2%). Its samples and the eight diagnostic profiles are separate and do not replace any sample in the full 103-query aggregate. Benchmark warmup and public documentation are committed in [f0a229f7b9](https://github.com/apache/doris/commit/f0a229f7b9) and pushed to [PR #69904](https://github.com/apache/doris/pull/69904). Full raw samples, before/after CSV/JSON, complete profiles and a reproduction report are preserved locally under `docs/delta-m1-preflight/benchmark/post-statistics-2026-11-02/`. The tasklist and design/acceptance records are updated; **63/67 remains the completed count and END-03 remains in progress**. All **1,227** MinIO objects match their original paths, sizes and SHA-256 before and after the run. FE/BE, the temporary HMS/PostgreSQL/MinIO services and Docker Desktop are stopped again; all **11** container objects, data and build artifacts are retained. -- 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]
