pure/docs/research/133-register-tiles-measured.md, branch masterVerified execution runtime unifying state, persistence, and replication.
http://git.yeowoo.org/pure/atom?h=master2026-10-08T13:36:27ZResearch 132 and 133 corrected: the per-batch cost was icount's three rounds over one round's batches, 3x too high; about 2,300 instructions (0.33 us) in a small block, 0.90 us in knapsack's (38% of its time, not 20%)2026-10-08T13:36:27Zeunhhunecpition205@gmail.com2026-10-08T13:36:27Zurn:sha1:3e3f008b942106cedb59ebca760376274c2fcb84Research 133: register tiles measured (132's deciding experiment): past the bar on roots (0.71x the compiled code's cycles a step), not on knapsack (4x) or BFS (2.4x), as NEON has no gather and a tile's lanes go out and in one by one; tight passes of the batch executor's own shape in the same run at 1.2x (BFS) and 0.4x (knapsack, clang): the shape is right and the 3x is implementation, so 132's A (a batch program made once at load, loops tight by type, width and stride, fewer passes) is the way; tiles kept aside for chains of arithmetic; 132, 024 and CHECKPOINT say so2026-10-08T12:39:18Zeunhhunecpition205@gmail.com2026-10-08T12:39:18Zurn:sha1:66e3f7d2adbc752b1e7a6c052675868a564314d4