Research 137: the batch executor's stage 4 (columns read in place, comparisons counting their lanes): knapsack 10.07 to 8.25 ms (bar 8.6; b9 11.46, the compiled code 7.9), its instructions 12.2% below b9's; what is left of knapsack's gap is the per-batch interpretation (about 3 ms), not the kernels (ADD 1.7, GT 1.7, the masked set 0.55 ms); the others the same by instructions; make check's server_view_workers timed out once (the known flake), passed 3/3 alone and in a full rerun; bank queries' instructions within 0.1% of b9's, TPC-C -0.3%; the two large gaps left (matmul's nested rounds, BFS's candidate tail) the user's to choose

1 files changed, 4 insertions, 0 deletions

diff --git a/research/experiments/040-against-c/2026-10-08-queries-s16-7.jsonl b/research/experiments/040-against-c/2026-10-08-queries-s16-7.jsonl
new file mode 100644
index 0000000..19e90f5
--- /dev/null
+++ b/research/experiments/040-against-c/2026-10-08-queries-s16-7.jsonl
@@ -0,0 +1,4 @@
+{"query": "per_branch", "rows": 10, "accounts": 100000, "pure_median_ms": 5.3, "pure_fastest_ms": 5.2, "sqlite_median_ms": 45.29, "sqlite_fastest_ms": 45.12, "sqlite_indexed_median_ms": 14.03, "sqlite_indexed_fastest_ms": 13.98}
+{"query": "in_credit", "rows": 1, "accounts": 100000, "pure_median_ms": 0.82, "pure_fastest_ms": 0.78, "sqlite_median_ms": 5.9, "sqlite_fastest_ms": 5.9, "sqlite_indexed_median_ms": 0.16, "sqlite_indexed_fastest_ms": 0.16}
+{"query": "large", "rows": 212, "accounts": 100000, "pure_median_ms": 0.41, "pure_fastest_ms": 0.36, "sqlite_median_ms": 1.49, "sqlite_fastest_ms": 1.47, "sqlite_indexed_median_ms": 1.5, "sqlite_indexed_fastest_ms": 1.48}
+{"query": "ranks", "rows": 404, "accounts": 100000, "pure_median_ms": 6.76, "pure_fastest_ms": 6.69, "sqlite_median_ms": 2845.34, "sqlite_fastest_ms": 2837.97, "sqlite_indexed_median_ms": 7.59, "sqlite_indexed_fastest_ms": 7.53}