benchmarks/sql: anti-DCE defense — accept() result in every bench run-block by borisbat · Pull Request #2756 · GaijinEntertainment/daScript

borisbat · 2026-05-20T14:34:38Z

Closing — combining with #2757 + Copilot fixes into a single PR rebased on fresh master (bbatkin/decs-zip-unroll-pattern3 already landed via #2750).

Will reopen as a single PR covering: m4_decs_fold lane + anti-DCE accept sweep + order_by+first splice arm + fix for first()-panic-on-empty semantics flagged by #2757 review.

Adds a third benchmark target alongside m1_sql / m3_array / m3f_array_fold: m4_decs_fold runs the same chain shape through `_fold(from_decs_template(type<DecsCar>)...)`. Gives a tri-platform comparison (SQL vs array vs decs) under one chain spec. Shared scaffold in _common.das: `[decs_template(prefix="car_")] DecsCar` mirroring Car's 6 fields + `fixture_decs(n)` parallel to `fixture_array(n)`. Each benchmark file gains a `run_m4` + `[benchmark]` wrapper. Lambda quirks: explicit `$(c : Car)` annotations don't match the decs tuple element type, so m4 lanes use `_select(_.field)` macro form (auto-types via macro expansion). first_or_default_match's sentinel is a named-tuple literal matching the iter element shape. Skipped (Cat C — need new decs surface): indexed_lookup (eid-lookup), join_count (decs join design), zip_dot_product (decs zip surface). Tracked in `benchmarks/sql/M4_DECS_EXPANSION.md` with full first-sweep results matrix, Cat A/B split, suspect-0ns-list, and Wave 2-4 plans (Cat C surface adds / Slice 5+ splice arms / per-chain component-narrowing perf). Wave 1 results (100K, INTERP): - Cat A m4 beats SQL on most aggregate/filter shapes (1.5-9x), ~3-5x slower than m3f due to 6-component multi-iter for-loop overhead even when chain reads one field - Cat B m4 falls to eager bridge (~100-130 ns); becomes the regression guard for each plan_decs_unroll splice arm as Slice 5+ lands Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

…-block Sweeps 197 b|>run blocks across 51 SQL benchmarks, inserting `b |> accept(<result_var>)` immediately after the inner let-binding. Uses the existing `[sideeffects]` helper `accept` from dastest/testing.das:172. **Why:** Documents intent (result must escape) and protects against future DCE if anyone modifies the chains. ast_dump verified the calls survive compilation: e.g. take_count m3f lowers to a full spliced invoke + accept(b, rows) + empty/failNow guard — the chain is genuinely running. **Finding:** This sweep was initially aimed at the 11 m3f=0 ns/op cases suspected of being DCE'd (select_count, take_count, take_count_filtered, take_sum_aggregate, reverse_take, skip_take, distinct_take, any_match, element_at_match, first_match, first_or_default_match). After the sweep, those cells still report 0 ns/op. Investigation: ast_dump shows the spliced loop body fully expanded and the accept call alive. The zeros are real — dastest reports total_time / n where n=100000, and a body cost ≤~100us divides to ≤1 ns/op and rounds to 0. For take(N)+to_array shapes the divisor is 100000 but only TAKE_N=1000 elements are processed, so the unit underreports actual per-element cost. Not a DCE artifact; matrix is honest. The accept guards stay regardless — cheap insurance for future bench changes, and the convention is uniform across all four lanes (m1/m3/m3f/m4). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Copilot

Pull request overview

This PR updates the SQL benchmark suite to make benchmark bodies harder to optimize away by ensuring each run-block “uses” its computed result via accept(...). It also expands the benchmark matrix with an m4_decs_fold lane (decs-backed) and documents the m4 expansion plan.

Changes:

Insert b |> accept(result) into benchmark run-blocks as an anti-DCE defense.
Introduce an m4_decs_fold benchmark lane across many SQL benchmarks (new fixture_decs + DecsCar template in _common.das, plus per-benchmark run_m4 + [benchmark] entrypoints).
Add documentation for the m4 lane expansion plan.

Reviewed changes

Copilot reviewed 53 out of 53 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
benchmarks/sql/_common.das	Add decs dependency + `DecsCar` template and `fixture_decs` to support m4 lane.
benchmarks/sql/aggregate_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/all_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/any_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/average_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/bare_order_where.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/chained_where.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/contains_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/count_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/distinct_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/distinct_take.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/element_at_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/first_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/first_or_default_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_average.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_first.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_having_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_having_hidden_sum.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_max.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_min.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_multi_reducer.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_select_sum.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_sum.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/groupby_where_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/groupby_where_sum.das	Add `accept(...)` + add m4 lane benchmark (note: nested `_select` issue flagged).
benchmarks/sql/indexed_lookup.das	Add `accept(...)` in benchmark loop.
benchmarks/sql/join_count.das	Add `accept(...)` in benchmark loop.
benchmarks/sql/last_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/long_count_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/max_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/min_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/order_take_desc.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/reverse_take.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/select_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/select_where.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/select_where_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/select_where_order_take.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/select_where_sum.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/single_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/skip_take.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/skip_while_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/sort_first.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/sort_take.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/sum_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/sum_where.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/take_count.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/take_count_filtered.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/take_sum_aggregate.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/take_while_match.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/to_array_filter.das	Add `accept(...)` + add m4 lane benchmark.
benchmarks/sql/zip_dot_product.das	Add `accept(...)` in benchmark loop.
benchmarks/sql/M4_DECS_EXPANSION.md	Add documentation for m4 lane expansion plan.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

+        let groups <- _fold(from_decs_template(type<DecsCar>)
+                            ._group_by(_.brand)
+                            ._select((Brand = _._0,
+                                      TotalPrice = _._1 |> _select(_.price) |> sum()))


+                            ._where(_.price > 500)
+                            ._group_by(_.brand)
+                            ._select((Brand = _._0,
+                                      TotalPrice = _._1 |> _select(_.price) |> sum()))


+        let groups <- _fold(from_decs_template(type<DecsCar>)
+                            ._group_by(_.brand)
+                            ._select((Brand = _._0,
+                                      AvgPrice = _._1 |> _select(_.price) |> average()))


+        let groups <- _fold(from_decs_template(type<DecsCar>)
+                            ._group_by(_.brand)
+                            ._select((Brand = _._0,
+                                      MinPrice = _._1 |> _select(_.price) |> min()))


+        let groups <- _fold(from_decs_template(type<DecsCar>)
+                            ._group_by(_.brand)
+                            ._select((Brand = _._0,
+                                      MaxPrice = _._1 |> _select(_.price) |> max()))


+                                      TotalPrice = _._1 |> _select(_.price) |> sum(),
+                                      MaxPrice   = _._1 |> _select(_.price) |> max()))


+    b |> run("m4_decs_fold/{n}", n) {
+        let groups <- _fold(from_decs_template(type<DecsCar>)
+                            ._group_by(_.brand)
+                            ._having(_._1 |> _select(_.price) |> sum > 50000)


+# m4_decs_fold lane — expansion plan
+
+Adds a third benchmark target alongside `m1_sql` (SQL via `_sql`) and `m3`/`m3f` (array linq, with/without `_fold` splice). The new lane is `m4_decs_fold`:
+
+```das
+_fold(from_decs_template(type<DecsCar>)._where(...)._select(...).sum())
+```
+
+Goal: tri-platform comparison (SQL vs array vs decs) under the same chain shape, so the splice path's decs win is directly comparable to its array win.


borisbat and others added 2 commits May 20, 2026 02:57

borisbat mentioned this pull request May 20, 2026

linq_fold: plan_order_family — order_by + first → min_by splice arm #2757

Closed

borisbat requested a review from Copilot May 20, 2026 14:45

Copilot started reviewing on behalf of borisbat May 20, 2026 14:46 View session

Copilot AI reviewed May 20, 2026

View reviewed changes

borisbat closed this May 20, 2026

borisbat mentioned this pull request May 20, 2026

linq_fold: m4_decs_fold bench lane + anti-DCE accept sweep + order_by+first splice arm #2760

Merged

7 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

benchmarks/sql: anti-DCE defense — accept() result in every bench run-block#2756

benchmarks/sql: anti-DCE defense — accept() result in every bench run-block#2756
borisbat wants to merge 2 commits into
bbatkin/decs-unroll-slice4from
bbatkin/bench-anti-dce

borisbat commented May 20, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		TotalPrice = _._1 \|> _select(_.price) \|> sum(),
		MaxPrice = _._1 \|> _select(_.price) \|> max()))

Conversation

borisbat commented May 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

borisbat commented May 20, 2026 •

edited

Loading