{"id":7028,"job_id":13835,"problem_id":6,"lane_id":34,"type":"measure","user_id":90,"model":"muse-spark","provider":"unknown","report_md":"# All-zeros measure: 60B il4 tally cell stays uniform; best 9\n\nJob 13835, measure, all-zeros. Improvement claim: none. The best\ncandidate reached score 9, below the 13 floor, so nothing goes to the\nresults page.\n\n## Hypothesis\n\nThe il4 tally cell is the track's honest baseline: its score tallies\nstay uniform at the 60B scale, and its loaded rate is stable across\ncycles. A fresh 60B counted cell from a verified disjoint nonce tests\nboth: tallies within about 2 sd of uniform and a rate matching the\nprior cycle's loaded 183.92 M/s would confirm the baseline, while any\ndrift would be a finding.\n\n## What was run\n\n- One 60B il4 tally cell, start nonce 12429840000000 (verified\n  disjoint: prior cell ran 12369840000000 + 60B), 1 thread, other core\n  running the parallel self-match cell. 185.00 M/s loaded. Best 9.\n- Tallies (score >= k): tally2 234,392,866 (uniform expectation\n  234,375,000, +1.17 sd), tally4 913,745 (expectation 915,527,\n  -1.86 sd), tally6 3,649 (expectation 3,576, +1.22 sd), tally7 226\n  (expectation 223.5, +0.17 sd), tally8 11 (expectation 14.0,\n  -0.79 sd). All within 2 sd: uniform. One score-9 hit vs 0.87\n  expected.\n- Best candidates, all hashlib-verified from the log (10\n  best-progression lines, 0 mismatches):\n  - score 9: digest 0000000005bfdfedda5c0cdfc8757e85\n  - score 8: digest 000000005cc6acaf62e75f49457eb489\n- Counted all-zeros total now about 1080B. Next disjoint all-zeros\n  start nonce: 12489840000000.\n\n## What this showed\n\nThe baseline holds: tallies uniform, rate stable (185.00 M/s loaded\nvs 183.92 M/s last cycle). Per the brief, this run is reported as a\nbaseline measurement: it used only the known il4 tool, and it found\nno structural shortcut.\n\n## Next run\n\nKeep extending the counted total with disjoint 60B il4 tally cells;\nperiodic uniformity controls stay cheap at this scale. The faster\nclang esfg6 arm (about 345 M/s solo) remains the production arm for\nbulk search, with il4 tally cells as controls.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":null,"status":"pending","final_rung":null,"created_at":"2026-11-06T12:46:12.682Z","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. Use the il4 tally binary built with gcc -O3 -march=znver4\n   -mtune=znver4 from the tally-patched source.\n2. Verify the next disjoint start nonce from the prior cell's done\n   line: start + total, never the notes number alone.\n3. Run the tally binary as md5tun_tally <start_nonce> 60000000000\n   120. Advance the start nonce by the full 60B for the next disjoint\n   cell: next disjoint start after this cycle is 12489840000000.\n4. Compare tallies against N x 16^-k for score >= k, k = 2, 4, 6, 7, 8.\n5. Verify every top hit by recomputing its digest with hashlib from\n   the logged input bytes before filing.","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":"2026-11-06T12:46:12.682Z","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 what makes the first output word of MD5 small, and use it to reach more leading zeros than generic search would with the same compute. Ideas to test: freedom from extra message blocks, neutral bits and message modification from collision attacks applied to the output instead of a difference, early abort on the final additions. 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/7028/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"d3a7e140dd5ca41a1763d55d4b4304a097ce6bba659c0e3fb4060328b36376f9","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"pending","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}