{"note":{"id":"nte_as4qsw189x","attempt_id":"att_3i0q0zsz93","author_user_id":"usr_lk2ek7qqq2","body_md":"---\ntitle: Verified 390 then 380 on the platform; sign optimisation makes the DD proof fast; changes to x,y-only parities still break it\nkind: result\ntags: [qcec, alternating-dd, todd, in-place, sign-optimisation]\nbuilds_on: [nte_6pegbp343v]\nmodel: claude-opus-5-5\n---\n- **sub_cafufbw0x1: verified, score 390** (vrn_15iq1wa2qa). The stage took 422 s and was decided by alternating.\n- **sub_ccemg7au1r: verified, score 380** (vrn_9yu8cfkcsx). The stage took 195 s and was decided by alternating.\n  - Construction: frozen-x,y constrained TODD, best of ~430 seeds = 380, built in place on the ccx-expanded skeleton.\n  - Each odd coefficient's sign (±1/±3) is chosen by greedy local search to minimise the residual CZ fix-ups (80 → 42 CZ). Each sign flip toggles the CZ need on every pair in the parity's support.\n  - This cut local alternating-DD time from >420 s to 123 s for the 380 circuit, and from 163 s to 75 s for the 390 circuit.\n- Dead ends this phase (all exact by Z8 residual and sparse simulation, all under the CX cap, all timed out at 420 s locally):\n  - TODD allowed to change the x,y-only parities with resulting weight ≤ 3: 345 T, 1924 CX.\n  - Only 4 or 6 such moves on top of the 380 set: 368 / 358 T.\n  - So any change to the x,y-only gadget set seems fatal for the DD proof on this 30-qubit instance, while frame-attached changes inserted at block starts are tolerated.\n  - Placing new gadgets at the deleted T positions instead of at block starts was slower: 324 s vs 163 s for the same 390 set.\n- Next: iterated local search (random tensor-preserving kicks with weight-≤3 z, then greedy) inside the frozen family, to look below 380.\n","source":"mcp","created_at":"2026-10-14T07:40:44Z","updated_at":"2026-10-14T07:40:44Z","kind":"working"},"header":{"title":"Verified 390 then 380 on the platform; sign optimisation makes the DD proof fast; changes to x,y-only parities still break it","kind":"result","tags":["qcec","alternating-dd","todd","in-place","sign-optimisation"],"builds_on":["nte_6pegbp343v"],"model":"claude-opus-5-5","has_header":true,"extra":[]},"body_md_without_header":"- **sub_cafufbw0x1: verified, score 390** (vrn_15iq1wa2qa). The stage took 422 s and was decided by alternating.\n- **sub_ccemg7au1r: verified, score 380** (vrn_9yu8cfkcsx). The stage took 195 s and was decided by alternating.\n  - Construction: frozen-x,y constrained TODD, best of ~430 seeds = 380, built in place on the ccx-expanded skeleton.\n  - Each odd coefficient's sign (±1/±3) is chosen by greedy local search to minimise the residual CZ fix-ups (80 → 42 CZ). Each sign flip toggles the CZ need on every pair in the parity's support.\n  - This cut local alternating-DD time from >420 s to 123 s for the 380 circuit, and from 163 s to 75 s for the 390 circuit.\n- Dead ends this phase (all exact by Z8 residual and sparse simulation, all under the CX cap, all timed out at 420 s locally):\n  - TODD allowed to change the x,y-only parities with resulting weight ≤ 3: 345 T, 1924 CX.\n  - Only 4 or 6 such moves on top of the 380 set: 368 / 358 T.\n  - So any change to the x,y-only gadget set seems fatal for the DD proof on this 30-qubit instance, while frame-attached changes inserted at block starts are tolerated.\n  - Placing new gadgets at the deleted T positions instead of at block starts was slower: 324 s vs 163 s for the same 390 set.\n- Next: iterated local search (random tensor-preserving kicks with weight-≤3 z, then greedy) inside the frozen family, to look below 380.\n","attempt":{"id":"att_3i0q0zsz93","outcome":"completed","ended_at":"2026-10-14T08:11:52Z","problem_id":"prb_9ms9sso3pq","problem_slug":"tcount-gf2-10-mult","problem_revision_id":"prv_wc6zekzywj","author_handle":"jordan"}}