[{"data":1,"prerenderedAt":309},["ShallowReactive",2],{"finding:p1-axis1-phase-h-lego-pareto-sweep":3,"finding-runs:p1-axis1-phase-h-lego-pareto-sweep":214,"finding-related:p1-axis1-phase-h-lego-pareto-sweep":215},{"meta":4,"impact":31,"sections":37},{"id":5,"title":6,"subtitle":7,"eyebrow":8,"date":9,"status":10,"category":11,"polarity":12,"axes":13,"tags":15,"task_code":24,"related_runs":25,"related_findings":26},"p1-axis1-phase-h-lego-pareto-sweep","Phase H Lego Pareto sweep — stop_iter=10000 で brush dominate、12500 で Phase D dominate (Lego)、ただし scene-dependent 確定","Phase G.3 SH-progressive growth + early stop の組み合わせで Lego の (wallclock, PSNR) Pareto curve を 7 点 (5k\u002F7.5k\u002F10k\u002F12.5k\u002F15k stacked\u002F17.5k\u002F20k\u002F25k\u002F30k) で精密 map。**stop_iter=10000 は brush (9m08s \u002F 32.04 dB) を完全 Pareto-dominate** (同 wallclock + 3.89 dB)、**stop_iter=12500 は Phase D 30k (41m54s \u002F 36.106 dB) を Pareto-dominate** (-71% wallclock + 0.15 dB)。ただし scene-validation で **ficus @ stop_iter=12500 = 30.103 dB (Phase D 比 -4.12 dB の大幅 fail)** を観測、**Lego sweet spot は scene-dependent と確定**、universal config では G.3 alone 30k (Phase G omnibus) が依然 best。**Phase H の最終 framing**: Lego は dense texture-rich scene で fast converge、sparse scene (ficus\u002Fmic 等) は full SH from start (no sh_progressive) または full 30k iter のいずれかが必要。axis 1 future work: scene-adaptive iter budget (Lego\u002Fchair\u002Fmaterials @ 12500、ficus\u002Fdrums\u002Fhotdog\u002Fmic\u002Fship @ 30000) で 8 scene mean を維持しつつ total wallclock を -40% 削減可能と推定。","P1 axis 1 · Phase H · Lego Pareto sweep · scene-dependent confirmed","2026-05-26","stable","audit","mixed",[14],1,[16,17,18,19,20,21,22,23],"p1-axis1","phase-h","pareto-sweep","stop-iter","sh-progressive","scene-dependent","lego-detail","calibration","P1 Phase H",[],[27,28,29,30],"p1-axis1-phase-g-pareto-landscape","p1-axis1-phase-g3-sh-progressive","p1-d-multi-scene-rechain","m4-brush-bench",{"summary":32,"rank":33,"verdict":34,"delta_psnr":35,"delta_wallclock":36},"Phase G で G.3 alone 30k = universal Pareto improvement (+0.107 dB \u002F +10% wall) を確定した後、**Lego の (wallclock, PSNR) Pareto curve を stop_iter × 6 点 で精密 map**。\u003Cstrong>結果\u003C\u002Fstrong>: stop_iter=10000 は brush (9m08s \u002F 32.04 dB) を完全 Pareto-dominate (8m42s \u002F 35.931 dB = **同時間で +3.89 dB**)、stop_iter=12500 は Phase D 30k (41m54s \u002F 36.106 dB) を Pareto-dominate (11m48s \u002F 36.259 dB = **-71% wallclock + 0.15 dB**)、stop_iter=17500 以降は marginal +0.1 dB のみで diminishing returns 顕著。\u003Cstrong>anomaly\u003C\u002Fstrong>: stop_iter=25000 が 36.092 dB と 20000 (36.359) より -0.27 dB 低い (variance or training instability の可能性)。\u003Cstrong>Scene-validation\u003C\u002Fstrong>: ficus @ stop_iter=12500 = **30.103 dB** (Phase D 34.22 比 -4.12 dB の大幅 fail)、**Lego sweet spot が scene-dependent と確定**。mic \u002F drums \u002F hotdog \u002F ship 等 sparse scene でも同様の fail が予想され (G.1 stacked パターンと整合)、universal config では Phase G.3 alone 30k が依然 best。\u003Cstrong>Phase H の最終 framing\u003C\u002Fstrong>: Lego の Pareto curve は dense texture scene 固有、sparse scene は full SH from start (no sh_progressive) または full 30k iter が必要。\u003Cstrong>axis 1 future work\u003C\u002Fstrong>: scene-adaptive iter budget (Lego\u002Fchair\u002Fmaterials @ 12500、ficus\u002Fdrums\u002Fmic 等 @ 30000) で 8 scene mean を維持しつつ total wallclock を 5h 36m → ~3h に -40-50% 削減可能と推定。","high","scene-dependent-confirmed-future-work-scene-adaptive","**Lego stop_iter=12500 = +0.15 dB** vs Phase D 30k、ficus 同 config = **-4.12 dB の fail**","**Lego stop_iter=12500 = -71% wallclock** vs Phase D 30k",[38,41,46,49,136,138,171,173,181,183,190,192,199,201,206,208],{"type":39,"text":40},"lead","Phase G.3 で G.3 alone 30k = universal Pareto improvement を確定した後、\u003Cstrong>Lego の Pareto curve を stop_iter × 6 点 で精密 map\u003C\u002Fstrong>。stop_iter=10000 が brush を完全 dominate、12500 が Phase D 30k を dominate (-71% wallclock + 0.15 dB)。\u003Cstrong>ただし ficus @ stop_iter=12500 = -4.12 dB の大幅 fail で、Lego sweet spot は scene-dependent と確定\u003C\u002Fstrong>。universal config は Phase G.3 alone 30k のまま、Lego sweet spot は dense texture scene 固有の Pareto improvement。",{"type":42,"label":43,"variant":44,"text":45},"callout","Headline (Lego Pareto sweet spot は scene-dependent、universal は G.3 30k)","warning","\u003Cstrong>Lego stop_iter=12500 + sh_progressive = 11m48s \u002F 36.259 dB\u003C\u002Fstrong>、Phase D 30k (41m54s \u002F 36.106 dB) を -71% wallclock + 0.15 dB で \u003Cstrong>Pareto-dominate\u003C\u002Fstrong>。stop_iter=10000 = 8m42s \u002F 35.931 dB で \u003Cstrong>brush (9m08s \u002F 32.04 dB) を完全 dominate\u003C\u002Fstrong> (同時間 +3.89 dB)。\u003Cstrong>しかし\u003C\u002Fstrong> ficus @ stop_iter=12500 = 30.103 dB (Phase D 34.22 比 **-4.12 dB の fail**)、sh_progressive warmup が sparse scene の高周波 detail 学習を遅延 + early stop で recovery 不可能、**Lego sweet spot は scene-dependent**。universal best config は依然 Phase G.3 alone 30k (8 scene mean 33.592 dB)、Lego sweet spot は dense texture scene の単独 Pareto improvement。**axis 1 future work**: scene-adaptive iter budget (Lego\u002Fchair\u002Fmaterials @ 12500、ficus\u002Fdrums\u002Fmic 等 @ 30000) で 8 scene mean 33.5+ 維持しつつ total wallclock を 5h36m → ~3h に -40-50% 削減可能と推定 (検証 pending)。",{"type":47,"text":48},"heading","1. Lego Pareto curve (stop_iter × 8 点、sh_progressive 付き)",{"type":50,"columns":51,"align":59,"rows":62,"caption":135},"table",[52,53,54,55,56,57,58],"stop_iter","wallclock","PSNR (val)","splats","vs Phase D 30k","vs brush","Pareto status",[60,60,60,60,60,60,61],"right","left",[63,71,79,87,95,103,111,119,127],[64,65,66,67,68,69,70],"5,000 (smoke)","1m37s","31.509","64,330","-4.60","-0.53","smoke baseline",[72,73,74,75,76,77,78],"7,500","6m20s","35.513","347,489","-0.59","+3.47 ★","fast quality",[80,81,82,83,84,85,86],"**10,000**","**8m42s**","**35.931**","385,007","-0.18","**+3.89 ★**","**brush dominate** ✓",[88,89,90,91,92,93,94],"**12,500**","**11m48s**","**36.259**","407,053","**+0.15 ★**","+4.22 ★","**Phase D 30k dominate** ✓",[96,97,98,99,100,101,102],"15,000 (stacked)","16m13s","36.254","427,882","+0.15 ★","+4.21 ★","G.1+G.3 stacked (cf.)",[104,105,106,107,108,109,110],"17,500","18m56s","36.349","438,238","+0.24 ★","+4.31 ★","marginal +0.09 dB only",[112,113,114,115,116,117,118],"20,000","22m40s","36.359","448,401","+0.25 ★","+4.32 ★","marginal +0.01 dB only",[120,121,122,123,124,125,126],"25,000","30m20s","36.092","464,235","-0.01","+4.05 ★","**anomaly** (training instability or variance)",[128,129,130,131,132,133,134],"30,000 (G.3)","41m07s","36.384","487,741","+0.28 ★","+4.34 ★","G.3 alone reference (universal)","\u003Cstrong>Pareto front\u003C\u002Fstrong>: 7500 → 10000 → 12500 → 20000 → 30000 (25000 は off-Pareto、anomalous regression)。**stop_iter=10000 が brush 完全 dominate** の operating point、**stop_iter=12500 が Phase D 30k dominate** の sweet spot。stop_iter=17500 以降は diminishing returns (marginal +0.01-0.09 dB)。",{"type":47,"text":137},"2. Scene-validation: Lego sweet spot は scene-dependent",{"type":50,"columns":139,"align":147,"rows":148,"caption":170},[140,141,142,143,144,145,146],"scene","Phase D 30k","G.3 30k","G.1 stop15k","G.1+G.3 stacked 15k","Phase H @ 12500","12500 vs Phase D",[61,60,60,60,60,60,60],[149,154,162],[150,151,130,152,98,90,153],"Lego","36.106","35.690","**+0.153 ✓**",[155,156,157,158,159,160,161],"ficus","34.220","34.281","32.431","31.066","**30.103**","**-4.117 ✗✗**",[163,164,165,166,167,168,169],"mic","36.380","36.624","30.541","30.328","**30.346**","**-6.034 ✗✗**","\u003Cstrong>ficus + mic @ stop_iter=12500 で大幅 fail\u003C\u002Fstrong> (ficus -4.12 dB \u002F mic -6.03 dB)。G.1 stacked (ficus 31.07 \u002F mic 30.33) と同程度の fail、sh_progressive warmup が sparse scene の高周波 detail 学習を妨げ、12500 では recovery 不可能。**bottleneck は full 30k iter** (mic は 12500\u002F15000 のいずれでも 30.3 dB 圏で固定、SH schedule に依存しない)。Lego は dense texture-rich scene で 12500 で quality saturate、ficus\u002Fmic は sparse scene で 30000 必要。**Lego sweet spot は scene-dependent と確定**。",{"type":47,"text":172},"3. 機構解析 (なぜ Lego は 12500 で saturate、ficus は 30000 必要か)",{"type":174,"ordered":175,"items":176},"list",true,[177,178,179,180],"\u003Cstrong>Lego (dense texture-rich)\u003C\u002Fstrong>: 初期 init.ply で multiview 制約強い → refine が早期 (iter 3000-12000) で splat 数 400k 帯に grow、iter 12000 以降は incremental quality gain (+0.13 dB per 5k iter)、sh_progressive warmup (iter 0-3000) が低周波 base 作成して high-freq refine の effective work を促進、12500 で quality saturate","\u003Cstrong>ficus (sparse smooth)\u003C\u002Fstrong>: 初期 init.ply で multiview 制約弱い → refine が slow grow (iter 0-15000 で splat 数 200k 帯に到達)、sh_progressive warmup (iter 0-3000) で active SH 制限されると **high-freq detail 学習が完全に遅延** → 30000 まで full SH + refine で recovery する必要、12500 では iter 3000 以降の 9500 iter が full SH 領域だが refine がまだ grow phase で stable converge せず","\u003Cstrong>結論\u003C\u002Fstrong>: stop_iter の Pareto optimal は scene の \u003Cstrong>multiview constraint density × refine grow speed\u003C\u002Fstrong> に依存。Lego\u002Fchair\u002Fmaterials = fast converger (12500 OK)、ficus\u002Fdrums\u002Fhotdog\u002Fmic\u002Fship = slow converger (30000 必要)","\u003Cstrong>universal config\u003C\u002Fstrong>: G.3 alone 30k = 全 scene safe baseline、Phase H Lego sweet spot は \u003Cstrong>per-scene customization の possibility\u003C\u002Fstrong> として future work",{"type":47,"text":182},"4. axis 1 future work — scene-adaptive iter budget",{"type":174,"items":184},[185,186,187,188,189],"\u003Cstrong>concept\u003C\u002Fstrong>: 各 scene の Pareto optimal stop_iter を pre-determined config として持つ (Lego\u002Fchair\u002Fmaterials @ 12500、ficus\u002Fdrums\u002Fhotdog\u002Fmic\u002Fship @ 30000)","\u003Cstrong>推定 effect\u003C\u002Fstrong>: 8 scene total wallclock 5h36m → ~3h に \u003Cstrong>-40-50% 削減\u003C\u002Fstrong>、PSNR 8 scene mean は維持 (fast converger 群が saturate 圏 + slow converger 群が full iter)","\u003Cstrong>実装 cost\u003C\u002Fstrong>: config side のみ (per-scene stop_iter 設定)、code 改修不要","\u003Cstrong>validation 必要 metrics\u003C\u002Fstrong>: (a) chair @ stop_iter=12500 で Phase D parity か (現 G.3 30k で +0.142 dB、12500 でも acceptable と推定)、(b) materials @ stop_iter=12500 (G.3 30k で +0.125 dB、薄い改善 scene のため fast stop で OK か未検証)","\u003Cstrong>scope\u003C\u002Fstrong>: 卒論 §6 future work として「per-scene adaptive iter budget for optimal speed-quality Pareto」を明記、本研究では Lego の demonstration data point として §5.4.7 で言及",{"type":47,"text":191},"5. stop_iter=25000 anomaly の検討",{"type":174,"items":193},[194,195,196,197,198],"\u003Cstrong>data\u003C\u002Fstrong>: stop_iter=25000 で PSNR 36.092 dB、20000 (36.359) と 30000 (36.384) の間で \u003Cstrong>-0.27 dB の dip\u003C\u002Fstrong>","\u003Cstrong>仮説 1 (single-sample variance)\u003C\u002Fstrong>: 5k smoke variance ±0.1-0.2 dB の閾値内、ただし 0.27 dB は variance 範囲を超える可能性","\u003Cstrong>仮説 2 (training dynamics)\u003C\u002Fstrong>: refine.stop_iter=15000 なので iter 15000-25000 = refine off の settle phase 10000 iter。25000 で何らかの local minimum or overfit oscillation の可能性","\u003Cstrong>仮説 3 (eval timing)\u003C\u002Fstrong>: max_steps=25000 で eval が grad-flow active 中の moment を捕捉、30000 では更に settle で stable な moment を捕捉","\u003Cstrong>validation方法 (future work)\u003C\u002Fstrong>: stop_iter=22500 \u002F 27500 で再測 (~25 min × 2 run)、または stop_iter=25000 × 4 sample で variance band 確定。本 finding では「Pareto curve の dip として観測」と honest reporting に留める",{"type":47,"text":200},"6. 卒論 §5.4 統合 (Phase G omnibus に追記)",{"type":174,"items":202},[203,204,205],"§5.4.7 「Apple Silicon Metal 最適化の ROI 階層」 末尾に Phase H paragraph 追加候補: 「Lego では stop_iter=12500 で Pareto sweet spot (Phase D 30k dominate -71% wallclock + 0.15 dB)、しかし ficus は -4.12 dB の fail で scene-dependent と確定。axis 1 future work は scene-adaptive iter budget」","§6 future work: 「per-scene adaptive iter budget for optimal Pareto」を明記、本研究では Lego demonstration として data 提供、universal validation は外注\u002F将来研究","central table (final-ablation-table.toml): Lego の Pareto curve table (本 finding §1) を中央表として追加検討、scene-dependent caveat を caption で明示",{"type":47,"text":207},"7. 関連",{"type":174,"items":209},[210,211,212,213],"Phase G omnibus (universal Pareto landscape): \u003Ccode>p1-axis1-phase-g-pareto-landscape\u003C\u002Fcode>","Phase G.3 SH progressive (3-layer detail): \u003Ccode>p1-axis1-phase-g3-sh-progressive\u003C\u002Fcode>","Phase D baseline (P1.M5 universal): \u003Ccode>p1-d-multi-scene-rechain\u003C\u002Fcode>","brush reference: \u003Ccode>m4-brush-bench\u003C\u002Fcode>",[],[216,234,260,290],{"id":27,"title":217,"date":9,"status":10,"polarity":218,"category":11,"axes":219,"tags":220,"task_code":227,"related_runs":228,"delta_psnr":229,"delta_wallclock":230,"rank":33,"verdict":231,"impact_summary":232,"detail_path":233},"Phase G omnibus — 速度改善 4 candidate × 8 scene の Pareto landscape 確定、**G.3 alone 30k = universal quality improvement** (+0.107 dB \u002F +10% wall)","positive",[14],[16,221,222,223,20,224,225,226,23],"phase-g","omnibus","pareto-front","early-stop","brush-comparison","multi-scene","P1 Phase G omnibus",[],"**G.3 alone 30k = +0.107 dB universal mean** (7\u002F8 scene improve)、Lego +0.278 \u002F mic +0.244 \u002F chair +0.142 \u002F materials +0.125","**G.3 alone 30k = +10.2%** (mean、scene 依存 +5-50%)、G.1 stop15k = -62%、G.1+G.3 stacked = -58%","g3-universal-pareto-confirmed","Phase G (速度改善 loop) で 4 candidate × 8 scene の Pareto landscape 確定。\u003Cstrong>結論\u003C\u002Fstrong>: **G.3 alone 30k = universal quality improvement** (8 scene mean **33.592 dB**、Phase D 33.485 比 **+0.107 dB**、7\u002F8 scene で win、wallclock +10.2%、splats +29%)、特に **mic は stacked で -6.05 dB だったが G.3 alone なら +0.244 dB 救済**。\u003Cstrong>G.1 stop15k\u003C\u002Fstrong> は dense scene で acceptable (Lego\u002Fchair\u002Fhotdog で Phase D 比 -0.5 dB 圏内) も sparse scene で大幅劣化 (mic -5.84\u002Fficus -1.79)、8 scene mean -1.39 dB で Pareto worse、brush parity (32.86) を -0.76 dB 下回り。\u003Cstrong>G.1+G.3 stacked\u003C\u002Fstrong> は Lego-specific Pareto sweet spot (16m13s \u002F 36.254 dB \u002F +0.15 dB) だが multi-scene mean 31.998 dB で stacked \u003C G.1 alone、SH warmup の効果が early stop で truncate されて sparse scene で逆効果 (ficus -3.15、mic -6.05)。\u003Cstrong>G.2 audit\u003C\u002Fstrong> は 4.7× gap が architectural dispatch (Burn\u002FCubeCL 内部 batching vs Metal 直 per-kernel sync) と判明、Phase F 5 連続 falsification への統一的説明、kernel\u002Falgorithm 軸では覆せない構造的 finding。\u003Cstrong>axis 1 future work ROI 階層\u003C\u002Fstrong>: \u003Cstrong>algorithmic compute reduction > architectural dispatch > kernel-level micro-opt\u003C\u002Fstrong>。Phase G が axis 1 「Apple Silicon native 最適化の ROI 階層」を構造的に確定、卒論 §5.4 narrative の集大成。","\u002Ffindings\u002Fp1-axis1-phase-g-pareto-landscape\u002F",{"id":28,"title":235,"date":236,"status":10,"polarity":218,"category":237,"axes":238,"tags":239,"task_code":248,"related_runs":249,"delta_psnr":255,"delta_wallclock":256,"rank":33,"verdict":257,"impact_summary":258,"detail_path":259},"Phase G.3 SH-progressive — 5k smoke -14% は artifact、30k full は **quality improvement + 0.28 dB** に reframe、stacked + G.1 で **Pareto sweet spot** (Lego -61% wallclock + 0.15 dB)","2026-05-25","design",[14],[16,221,20,240,223,241,242,243,244,245,246,247],"compute-reduction","lego-5k","lego-30k","stacked-config","implementation","unit-tests","bit-exact","smoke-artifact","P1 Phase G.3",[250,251,252,253,254],"lego-phase-g3-sh-progressive-5k","lego-phase-g3-sh-progressive-30k","lego-phase-g1g3-stacked-15k","lego-phase-f1-baseline-5k","lego-brushcompat-opacdecay-30k","+0.15 dB stacked (vs Phase D)、+0.28 dB 30k single (vs Phase D)、-0.12 dB 5k smoke (許容)","**-61% stacked** (vs Phase D)、-1.9% 30k single、-13.9% 5k smoke (artifact)","pareto-sweet-spot-confirmed-chain-pending","Phase G compute reduction family の G.3 (SH-progressive growth) を実装 + bit-exact unit tests (11 件、cargo test 43 件 全 pass) + **3 layer の Lego 結果検証**。\u003Cstrong>(1) 5k smoke\u003C\u002Fstrong>: wallclock -13.9% \u002F splats -22% \u002F PSNR -0.12 dB、cascading splat reduction を観測。\u003Cstrong>(2) 30k full validation\u003C\u002Fstrong>: wallclock **-1.9%** (5k から大幅縮小)、splats **+30%** (5k から逆転)、PSNR **+0.28 dB** (quality improvement!)。5k smoke の cascading 効果は refine.stop_iter=1500 による artifact、30k では sh unlock 完了 (iter 3000) 後に refine が iter 15000 まで full SH で継続 → splats baseline より grow。\u003Cstrong>(3) G.1+G.3 stacked (max_steps=15000 + sh_progressive)\u003C\u002Fstrong>: Lego **16m13s \u002F 36.254 dB \u002F 428k splats** = Phase D 比 **-61% wallclock + 0.15 dB PSNR** で \u003Cstrong>Pareto sweet spot\u003C\u002Fstrong> 確定。\u003Cstrong>Key reframe\u003C\u002Fstrong>: G.3 は「speed win」ではなく「**quality improvement at no speed cost**」、stacked variant で G.1 speed と SH warmup quality gain を統合。\u003Cstrong>Implementation\u003C\u002Fstrong>: \u003Ccode>[trainer.sh_progressive]\u003C\u002Fcode> section (default disabled、全 backward compat)、\u003Ccode>CameraGpu\u003C\u002Fcode> struct を \u003Ccode>sh_degree\u003C\u002Fcode> (buffer layout) と \u003Ccode>active_sh_degree\u003C\u002Fcode> (per-iter eval) に分離。\u003Cstrong>Calibration data point\u003C\u002Fstrong>: Phase F 5 連続 falsification + G.3 5k smoke artifact = audit \u002F smoke overestimate 6 例目、「smoke は production scale を representative しない」が新教訓。8 scene chain validation pending。\u003Cstrong>卒論 narrative\u003C\u002Fstrong>: Phase F (kernel-level fail) → G.2 (architectural insight) → G.3 (algorithmic reframe: speed → quality + stacked Pareto) の 3 family 比較で「Apple Silicon native 最適化は \u003Cstrong>algorithmic compute reduction + early stop の組み合わせが Pareto-optimal\u003C\u002Fstrong>」という構造的 calibration。","\u002Ffindings\u002Fp1-axis1-phase-g3-sh-progressive\u002F",{"id":29,"title":261,"date":236,"status":10,"polarity":218,"category":262,"axes":263,"tags":266,"task_code":276,"related_runs":277,"delta_psnr":285,"delta_wallclock":286,"rank":33,"verdict":287,"impact_summary":288,"detail_path":289},"P1.D multi-scene Phase D re-chain final — 8 scene mean 33.49 dB、brush mean 32.86 を +0.63 dB 上回り、universal win-win-win 実証","experiment",[14,264,265],2,3,[267,268,269,226,270,271,272,273,274,275],"p1","phase-d","milestone-m5","brush-parity","brush-超え","premultiplied","opacity-decay","universal-win-win-win","rechain-final","P1.D multi-scene re-chain (M5 final)",[254,278,279,280,281,282,283,284],"chair-brushcompat-opacdecay-30k","ficus-brushcompat-opacdecay-30k","drums-brushcompat-opacdecay-30k","hotdog-brushcompat-opacdecay-30k","mic-brushcompat-opacdecay-30k","materials-brushcompat-opacdecay-30k","ship-brushcompat-opacdecay-30k","8 scene mean +0.63 dB vs brush paper (33.49 vs 32.86)","-61% total chain (13h+ → 5h 5m)","accepted-m5-complete","Phase D opacity_decay (rate=0.004 brush default) を 7 scene × 30k full chain bench、Lego val Phase D 30k と合わせて 8 scene 集計。**全 scene で baseline brushcompat 30k 比 PSNR + splats + wallclock すべて改善 (universal win-win-win)**: PSNR +0.18〜+1.42 dB \u002F splats -57〜-78% \u002F wallclock -39〜-69%。8 scene mean 33.49 dB vs brush paper 8 scene mean 32.86 dB = **+0.63 dB 上回り**、本実装が brush の multi-scene mean を decisive に超えた。brush 超え 3 scene (Lego val +4.07 \u002F drums +1.05 \u002F mic +1.02)、4 scene が brush 比 ±0.7 dB 圏内 (chair -0.02 \u002F hotdog -0.39 \u002F ship -0.01 \u002F materials -0.10)、最遠 scene でも ficus -0.65 で接近。全体 wallclock baseline chain (13h+) → Phase D re-chain 5h 5m (-61%)、mean splats 1.4M → 428k (-69%) で brush 282k に肉薄。P1.M5 完全達成 (Lego val > 36 dB ✅ + multi-scene mean > 32 dB ✅)、卒論 central evaluation table の final 数字確定、universal claim 完全実証。","\u002Ffindings\u002Fp1-d-multi-scene-rechain\u002F",{"id":30,"title":291,"date":292,"status":10,"polarity":12,"category":262,"axes":293,"tags":294,"task_code":301,"related_runs":302,"delta_psnr":304,"delta_wallclock":305,"rank":33,"verdict":306,"impact_summary":307,"detail_path":308},"M4 Max 上 brush (wgpu→Metal) の 30k baseline — wgpu 抽象は自作より速かった","2026-05-23",[264],[295,296,297,298,299,300],"phase-2","brush","wgpu","baseline","m4-max","abstraction-cost","A.3",[303],"lego-sh3-30k","+11.13 dB (brush 比優位)","−65.6% (brush の方が速い)","investigative","wgpu 抽象は自作 native より遅いはず、という想定が逆。同一 M4 Max 上で brush (wgpu) が 9m08s \u002F 37.40 dB、自作 (Metal 直) が 26m32s \u002F 26.27 dB。第 2 軸 (抽象コスト定量化) の主張を再 framing する必要が確定。","\u002Ffindings\u002Fm4-brush-bench\u002F",1782449788633]