mirror/BFM-decomp
1
0
Fork 0
mirror of https://github.com/Druthulu/BFM-decomp synced 2026-09-27 05:56:00 -04:00
Code Issues Packages Projects Releases Wiki Activity
Files
16a1dabc798db0eefe5d07de3bf0bf7b5a9d8f8d
BFM-decomp/docs/progress.fleet.md
T
Drew T 1859266d60 feat(phase-30 S10): Sonnet wave — 13 heads + 57 members; the §136i ~120 boundary is too LOW 
- DREW'S CALL (2026-08-03): route the 30-target x2-9 wave to SONNET instead of Opus. The §136i
  >=120-ins Opus threshold was MY EXTRAPOLATION, never measured; this wave (125-793 ins) probes
  exactly the region where there was no data.
- RESULT: 13 banked of 16 that ran = **81%**, vs Opus's 10/13 = 77% on the comparable S8-3 slice.
  At least 8 banked SONNET-DIRECT (65 agents spawned: 60 sonnet, 5 opus escalations). Propagated
  57 member-matches / 4 failed across 42 overlays. 70 instances. R22 clean-fleet 140/140.
  FLEET 96.01% fn / 93.6% instr / 88.0% distinct (77,550 uniq).
  => **Sonnet is at least as capable as Opus on 125-793 ins. The ~120 boundary is too low.**
  NOT rewriting it to a specific number yet: 16 samples under a throttle confound cannot name a
  cliff. The controlled A/B (task #12) is how that number gets fixed properly.
- THE REAL LIMITER IS CAPACITY, NOT CAPABILITY: 14 of 30 agents were killed by SERVER-side
  throttling ("Server is temporarily limiting requests (not your usage limit)") that 30 concurrent
  Opus agents did not trigger. Practical rule: run Sonnet waves at ~12-16 concurrency, not 30.
  The 14 unrun targets are listed in the checkpoint for a smaller-batch retry.
- Sonnet's work quality was not shallow — three examples: func_8018797C read local-alloc.c and
  forced loads into an AGGREGATE to stop find_free_reg greedily taking 3 callee-saved regs;
  func_8018D870 used §136c sibling-first for ~70% of the body then blocked a coalesce with a pin;
  func_8017E3AC diagnosed an RC-3 callee-saved-order swap and noted the pin must be s32 or a stray
  `andi 0xffff` appears.
- MY ERROR, RECORDED: `until [ -s <output> ]` fires at the FIRST LINE of output, not at completion.
  It fired mid-propagation and I ran `make clean` on top of a live family_sweep, deleting asm/ and
  aborting both the regen and the sweep (corpus's R32 assertion refused to answer rather than return
  a wrong stub set — working as designed). No bad bytes: R22 verified 140/140 immediately after, and
  the propagation simply re-ran clean. Correct waiter is `pgrep -x make` (exact process name), which
  also cannot self-match the way `pgrep -f <pattern>` did when it leaked 4 waiter shells earlier.
  Third instance today of ONE root cause: trusting a proxy instead of the thing itself (a weight
  column vs a probe §136h; an exit status vs build output §136a; file-existence vs process exit).
2026-08-03 22:27:19 -06:00

9.1 KiB
Raw Blame History

BFM FLEET matching progress (generated by tools/progress.py --fleet — authoritative)

140 binaries: main + resident + 138 location overlays. DETERMINISTIC,

source-derived (committed src/*.c + config/dedup.us.yaml). Live byte gate: make check-all;

cross-binary collapsible-byte leverage: docs/duplicates.cross.md.

THREE progress metrics (all matter — see the labels):

FLEET fn-count byte-ident: 339609 / 353721 = 96.01% (REAL+LINKED+empties; FUNCTION-count, ×134-inflated — one crack counts per overlay) FLEET instr-weighted : 12305370 / 13141652 = 93.6% (shipped .text across main + resident + 138 overlays; the decomp.dev-DISPLAY number) FLEET distinct-code(uniq): 4961017 / 5634875 = 88.0% (77550/87459 unique fns; the DISTINCT-RE number) MAIN game-code weighted : 436 / 60201 = 0.7% (INCLUDED in the fleet numbers above since 2026-07-22 — roadmap §1 metrics contract; LINKED-excluding Ghidra sig dated 2026-06-14; caveat is R34: no independent second oracle for a PS-X EXE, NOT drift) (fleet EXCLUDING main, for continuity with pre-2026-07-22 readings: 12304934 / 13081451 = 94.1%)

FLEET REAL substantive : 337754 (of which dedup-shared 240422 via 1905 groups / 240496 instances) FLEET LINKED PsyQ objs : 959 FLEET NON_MATCHING : 7 (0 in any default build — G4) FLEET INCLUDE_ASM stubs : 14105 FLEET matchable : 353721

binary REAL shared LINKED byte-ident matchable byte-ident %
main 54 2 959 1055 2096 50.3%
resident 129 0 0 131 145 90.3%
ov_SC01_000 2377 1748 0 2377 2403 98.9%
ov_SC01_001 2385 1749 0 2387 2466 96.8%
ov_SC01_004 2379 1740 0 2380 2414 98.6%
ov_SC01_005 2409 1763 0 2409 2503 96.2%
ov_SC01_006 2409 1763 0 2409 2503 96.2%
ov_SC01_008 2379 1740 0 2381 2426 98.1%
ov_SC01_009 2405 1741 0 2406 2507 96.0%
ov_SC01_074 2384 1741 0 2386 2425 98.4%
ov_SC01_077 2536 1707 0 2538 2586 98.1%
ov_SC01_080 2412 1744 0 2412 2512 96.0%
ov_SC01_084 2437 1749 0 2442 2579 94.7%
ov_SC02_000 2502 1788 0 2502 2683 93.3%
ov_SC02_003 2502 1788 0 2502 2683 93.3%
ov_SC02_004 2377 1744 0 2377 2401 99.0%
ov_SC02_005 2524 1740 0 2534 2927 86.6%
ov_SC02_011 2607 1747 0 2618 2893 90.5%
ov_SC02_015 2388 1746 0 2388 2414 98.9%
ov_SC02_016 2469 1746 0 2472 2545 97.1%
ov_SC02_017 2537 1746 0 2545 2732 93.2%
ov_SC02_021 2393 1746 0 2393 2437 98.2%
ov_SC02_026 2484 1744 0 2490 2572 96.8%
ov_SC02_027 2541 1744 0 2550 2692 94.7%
ov_SC02_028 2546 1744 0 2556 2701 94.6%
ov_SC02_031 2480 1745 0 2485 2562 97.0%
ov_SC02_035 2468 1744 0 2471 2520 98.1%
ov_SC02_039 2382 1744 0 2382 2416 98.6%
ov_SC02_041 2459 1744 0 2462 2561 96.1%
ov_SC03_001 2645 1745 0 2662 2869 92.8%
ov_SC03_002 2521 1749 0 2536 2631 96.4%
ov_SC03_003 2386 1744 0 2387 2423 98.5%
ov_SC03_006 2535 1750 0 2544 2767 91.9%
ov_SC03_007 2449 1744 0 2453 2623 93.5%
ov_SC03_010 2402 1744 0 2402 2471 97.2%
ov_SC03_011 2404 1744 0 2410 2527 95.4%
ov_SC03_012 2376 1744 0 2377 2406 98.8%
ov_SC03_013 2408 1744 0 2408 2493 96.6%
ov_SC03_014 2555 1772 0 2555 2685 95.2%
ov_SC03_015 2555 1772 0 2555 2685 95.2%
ov_SC03_023 2388 1744 0 2389 2435 98.1%
ov_SC03_024 2497 1749 0 2504 2641 94.8%
ov_SC03_028 2445 1744 0 2449 2664 91.9%
ov_SC03_029 2488 1746 0 2498 2644 94.5%
ov_SC03_030 2406 1749 0 2408 2495 96.5%
ov_SC03_031 2435 1744 0 2438 2514 97.0%
ov_SC03_089 2493 1749 0 2500 2581 96.9%
ov_SC03_090 2503 1749 0 2511 2625 95.7%
ov_SC03_091 2517 1744 0 2525 2640 95.6%
ov_SC03_092 2486 1749 0 2498 2587 96.6%
ov_SC03_093 2500 1744 0 2504 2563 97.7%
ov_SC03_094 2484 1744 0 2490 2577 96.6%
ov_SC03_095 2416 1744 0 2419 2474 97.8%
ov_SC03_096 2413 1744 0 2416 2465 98.0%
ov_SC03_097 2507 1744 0 2514 2603 96.6%
ov_SC03_098 2489 1744 0 2492 2541 98.1%
ov_SC03_099 2470 1744 0 2473 2504 98.8%
ov_SC03_100 2505 1744 0 2509 2539 98.8%
ov_SC03_101 2465 1749 0 2469 2528 97.7%
ov_SC03_102 2454 1744 0 2457 2492 98.6%
ov_SC03_103 2477 1746 0 2480 2510 98.8%
ov_SC03_104 2522 1749 0 2529 2616 96.7%
ov_SC03_105 2430 1744 0 2437 2597 93.8%
ov_SC03_108 2385 1744 0 2385 2443 97.6%
ov_SC03_109 2395 1744 0 2397 2424 98.9%
ov_SC03_110 2427 1744 0 2427 2468 98.3%
ov_SC03_111 2457 1744 0 2460 2510 98.0%
ov_SC03_112 2418 1746 0 2420 2526 95.8%
ov_SC03_113 2419 1746 0 2422 2468 98.1%
ov_SC03_114 2377 1744 0 2379 2415 98.5%
ov_SC03_115 2418 1744 0 2420 2472 97.9%
ov_SC03_116 2385 1744 0 2388 2438 97.9%
ov_SC03_117 2453 1744 0 2459 2557 96.2%
ov_SC03_118 2523 1768 0 2524 2685 94.0%
ov_SC03_119 2519 1768 0 2520 2685 93.9%
ov_SC03_121 2391 1744 0 2394 2459 97.4%
ov_SC03_124 2555 1740 0 2575 2741 93.9%
ov_SC03_125 2508 1744 0 2520 2588 97.4%
ov_SC03_126 2387 1746 0 2387 2423 98.5%
ov_SC04_000 2460 1748 0 2469 2546 97.0%
ov_SC04_002 2487 1744 0 2491 2636 94.5%
ov_SC04_003 2424 1744 0 2428 2502 97.0%
ov_SC04_004 2457 1744 0 2459 2558 96.1%
ov_SC04_005 2481 1744 0 2487 2612 95.2%
ov_SC04_006 2417 1744 0 2419 2454 98.6%
ov_SC04_007 2475 1744 0 2479 2569 96.5%
ov_SC04_008 2388 1744 0 2388 2415 98.9%
ov_SC04_009 2414 1744 0 2417 2441 99.0%
ov_SC04_010 2390 1744 0 2391 2419 98.8%
ov_SC04_011 2501 1749 0 2507 2803 89.4%
ov_SC04_012 2385 1744 0 2386 2420 98.6%
ov_SC04_015 2513 1745 0 2524 2611 96.7%
ov_SC04_016 2388 1744 0 2390 2440 98.0%
ov_SC04_018 2657 1780 0 2657 2857 93.0%
ov_SC04_019 2649 1780 0 2649 2857 92.7%
ov_SC04_020 2495 1744 0 2507 2567 97.7%
ov_SC04_021 2387 1746 0 2387 2423 98.5%
ov_SC05_000 2394 1748 0 2397 2422 99.0%
ov_SC05_001 2462 1744 0 2467 2574 95.8%
ov_SC05_002 2417 1744 0 2420 2442 99.1%
ov_SC05_003 2398 1744 0 2399 2481 96.7%
ov_SC05_004 2408 1744 0 2410 2464 97.8%
ov_SC05_005 2424 1749 0 2425 2491 97.4%
ov_SC05_006 2410 1744 0 2410 2430 99.2%
ov_SC05_007 2417 1744 0 2422 2482 97.6%
ov_SC05_008 2430 1744 0 2432 2543 95.6%
ov_SC05_009 2395 1744 0 2399 2438 98.4%
ov_SC05_010 2433 1744 0 2437 2588 94.2%
ov_SC05_011 2384 1744 0 2385 2409 99.0%
ov_SC05_017 2628 1741 0 2639 2842 92.9%
ov_SC05_018 2524 1744 0 2537 2673 94.9%
ov_SC05_019 2387 1746 0 2387 2423 98.5%
ov_SC06_000 2511 1753 0 2514 2691 93.4%
ov_SC06_006 2409 1744 0 2410 2511 96.0%
ov_SC06_008 2463 1746 0 2469 2542 97.1%
ov_SC06_010 2437 1744 0 2442 2517 97.0%
ov_SC06_011 2401 1744 0 2405 2468 97.4%
ov_SC06_013 2391 1744 0 2392 2425 98.6%
ov_SC06_014 2402 1744 0 2404 2453 98.0%
ov_SC06_015 2396 1744 0 2396 2421 99.0%
ov_SC06_016 2438 1744 0 2440 2549 95.7%
ov_SC06_018 2464 1746 0 2471 2665 92.7%
ov_SC06_020 2425 1746 0 2426 2518 96.3%
ov_SC06_022 2451 1744 0 2459 2642 93.1%
ov_SC06_024 2470 1744 0 2476 2667 92.8%
ov_SC06_025 2439 1744 0 2444 2572 95.0%
ov_SC06_027 2379 1744 0 2380 2408 98.8%
ov_SC06_029 2460 1744 0 2472 2662 92.9%
ov_SC06_030 2399 1744 0 2399 2456 97.7%
ov_SC06_032 2457 1745 0 2464 2658 92.7%
ov_SC06_033 2456 1745 0 2463 2631 93.6%
ov_SC07_000 2419 1748 0 2421 2521 96.0%
ov_SC07_001 2391 1744 0 2393 2454 97.5%
ov_SC07_002 2444 1744 0 2448 2579 94.9%
ov_SC07_006 2296 1560 0 2376 2456 96.7%
ov_SC07_007 2313 1592 0 2397 2613 91.7%
ov_SC07_008 2373 1744 0 2373 2386 99.5%
ov_SC07_009 2385 1744 0 2387 2430 98.2%
ov_SC07_010 2295 1614 0 2378 2524 94.2%
ov_SC07_011 2314 1591 0 2394 2449 97.8%
Reference in New Issue View Git Blame Copy Permalink