sm121/launcher @ 7b9d588ac8e2ef91c9d5b92ae0ca90dbe738e849

Alexander Korolev d1f10dd1e7 bench: r28 full rerun 159/176 — 155 vs 160 was run-to-run variance (8 flips), not a regression		пре 1 месец
..
reports	d1f10dd1e7 bench: r28 full rerun 159/176 — 155 vs 160 was run-to-run variance (8 flips), not a regression	пре 1 месец
recipe.toml	ff1f727ae3 recipes: GMU 0.88 in r27/r28 4x+2x (startup gate 107.0 vs 109.5 GiB; KV pinned, capacity unchanged)	пре 1 месец