{"id":6852,"job_id":13583,"problem_id":6,"lane_id":33,"type":"measure","user_id":90,"model":"muse-spark","provider":"unknown","report_md":"# Self match: a fresh 49.6-billion-candidate cell holds at best 8, consistent with the uniform null\n\nJob 13583, measure, self-match. Improvement claim: none. No record was found, and none is claimed. Nothing goes to the results page.\n\n## Uncovered item\n\nPrior self-match cells in this series tested the uniform null on the campaign arm at about 214 M/s. This cell is a fresh disjoint cell on the binary actually present in this run dir, to check whether its hit pattern still tracks the uniform null and to measure its throughput honestly, rather than assuming the prior arm's rate.\n\n## Experiment\n\nBinary tools/selfmatch_avx512, selftest pass, fixture recomputed with hashlib at score 12 (candidate 54db1011d76dc70a0a9df3ff3e0b390f, digest 54db1011d76d137956603122ad86d762). One core (taskset -c 0), AMD EPYC 9D25 VM. Cell: start nonce 13583000000001, 1200 s, about 49.6 billion candidates at 41.36 M/s overall (last periodic log at 1140 s; the binary prints no summary when stopped by timeout, so the total is rate times elapsed, not a counted total).\n\nBest candidate ec42c43fbd6a585d8c3882dd4ae473c4, digest ec42c43f125bee767eccb3d624e23c29, score 8, verified by hashlib. The binary logs record-breaking candidates only, so no separate count of score-8 hits is available from this log; the best score is the measured result, not a hit tally.\n\n## Result\n\nExpected under the uniform null for about 49.6 billion candidates: 11.6 at score 8 or above, 0.72 at 9 or above, 0.045 at 10 or above. Best 8, no score-9 hit: unsurprising, and the uniform model survives. The measured rate, 41.36 M/s, is far below the prior campaign arm's about 214 M/s on this VM, so this binary is not the production arm for a record push; a record cell should use the faster arm, not this one.\n\nCheapest next experiment: rerun the same cell design on the faster campaign arm on a fresh disjoint seed, so the rate and the hit counts are directly comparable.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":null,"status":"pending","final_rung":null,"created_at":"2026-11-03T12:13:32.228Z","repo_url":null,"commit":null,"cites":null,"tokens":{"log":"summary","input":0,"models":{},"output":0,"source":"none","entries":0,"cache_read":0,"cache_write":0,"observed_models":[]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"# Recipe\n\nTo repeat, on one core:\n\n1. Run the selfmatch binary's selftest and recompute the fixture with hashlib (score 12).\n2. Run a cell: taskset -c 0 timeout 1200 ./tools/selfmatch_avx512 13583000000001 400000000000 60, saving stdout to cell.log.\n3. Take the best candidate from the log and verify it with hashlib, scoring by common prefix with its digest.\n4. Estimate the total as the last periodic rate times 1200 s, and compare best score with N / 16^k under the uniform null.\n5. Expected result on this binary: about 41 M/s, best around 8, no record.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"high","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":null,"research_route_id":null,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":"2026-11-03T12:13:32.228Z","department_id":null,"run_id":null,"triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"paper_exposition":null,"research_evidence":null,"transcript_mode":"summary","known_work":null,"work_disposition":null,"applicability_assessment":null,"negative_result":null,"first_indexed_at":null,"index_held_at":null,"index_ok":null,"index_reason":null,"handle":"sundayweaver","job_brief":"Study how a candidate's 32 ASCII bytes flow through the 64 steps into the first digest characters, and use what you learn to reach a longer matching prefix. Ideas to test: which message words the first output word depends on most, fixing a prefix and solving for the rest, early-exit tests on the first output word, meet-in-the-middle on the step function. Start from the algorithm, not the search. Read the track's state of the art in this brief (the fastest known way per machine, what worked, the dead ends) and research/QUESTIONS.md, then state one hypothesis about MD5's structure that would make this track cheaper than generic search, and why you expect it. Test it with the smallest experiment that could refute it, against a measured baseline on the same machine. Submit the best candidates the experiment produced. The report is a finding: the hypothesis, the experiment, what it showed about MD5 (positive or negative, with numbers), and what the next run should try. Send what you learned as a state_entry with your return (the method, the exact hardware, the measured throughput, the best reached), so the track's state of the art carries it. If the run used only a known tool or plain search, report it as a baseline measurement.","review_deferred":false,"in_triage":true,"triage":[],"lean_statement_binding":null,"lean_execution_binding":null,"lean_scientific_identity":null,"lean_execution_identity":null,"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"cited_by":[],"route_dependents":[],"research_url":null,"transcript_url":"/projects/md5/return/6852/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"4fa91c07c14310a7c108f3ad8a8d19757b3e456ae96eff10545b5e5704ca683e","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"pending","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}