[{"data":1,"prerenderedAt":169},["ShallowReactive",2],{"run:s2-12500-noprof-r3":3,"run-findings:s2-12500-noprof-r3":136},{"run":4,"config":18,"metrics":81,"curve":87,"assets":135},{"id":5,"title":5,"subtitle":6,"eyebrow":7,"date":8,"workspace":9,"commit":10,"tags":11,"verdict":17},"s2-12500-noprof-r3","S2 gate: forward 1 CB 化 (chained forward + hierarchical GPU scan) の 12.5k full run repeat 3","Run summary · 2026-08 1× 工程表 v2 — AC 復帰後 計測 queue #2 (S2 gate)","2026-08-16","splat","66f7f34",[12,13,14,15,16],"p2-s2-forward-chained","s2","gpu-gate","no-profile","lego-12500","partial",[19,22,25,28,31,34,37,40,43,46,49,52,55,58,61,64,67,70,73,76,78],{"key":20,"value":21},"dataset","\u002FUsers\u002Fotkrickey\u002Fdev\u002F3dgs-workspace\u002Fdatasets\u002Fnerf_synthetic\u002Flego",{"key":23,"value":24},"gt_convention","Premultiplied",{"key":26,"value":27},"iterations","12,500",{"key":29,"value":30},"seed","42",{"key":32,"value":33},"capacity","1,000,000 splats",{"key":35,"value":36},"sh_degree","3",{"key":38,"value":39},"loss","L1Ssim",{"key":41,"value":42},"lambda","0.200",{"key":44,"value":45},"ssim","window=7 sigma=1",{"key":47,"value":48},"backend.backward","Simd",{"key":50,"value":51},"backend.loss_path","Gpu",{"key":53,"value":54},"backend.emit_pairs","Original",{"key":56,"value":57},"backend.forward","F32",{"key":59,"value":60},"backend.dispatch","PerStage",{"key":62,"value":63},"backend.radix_scan","GpuHierarchical",{"key":65,"value":66},"backend.binning","Resident",{"key":68,"value":69},"backend.tile_sort","GlobalRadix",{"key":71,"value":72},"backend.submission","Chained",{"key":74,"value":75},"backend.target_cache","true",{"key":77,"value":75},"backend.shader_fast_math",{"key":79,"value":80},"os_build","26.5.2 (25F84)",{"psnr":82,"wallclock":83,"wallclock_regress":84,"splats":85,"final_loss":86},36.15475082397461,"3m 1s",false,409861,"5.516684e-3",{"loss":88,"splats":119},{"iters":89,"values":104},[90,91,92,93,94,95,96,97,98,99,100,101,102,103],1,1000,2000,3000,4000,5000,6000,7000,8000,9000,10000,11000,12000,12500,[105,106,107,108,109,110,111,112,113,114,115,116,117,118],0.5938446521759033,0.05435022711753845,0.01370299980044365,0.008371880277991295,0.00748113077133894,0.006707792170345783,0.0062702251598238945,0.005993512459099293,0.005843691993504763,0.005691649857908487,0.005598301067948341,0.005565616302192211,0.005496924743056297,0.005516683682799339,{"iters":120,"values":121},[90,91,92,93,94,95,96,97,98,99,100,101,102,103],[122,123,124,125,126,127,128,129,130,131,132,133,134,85],5207,8223,182428,297700,349602,375282,389849,398364,403312,405650,406828,407813,409091,{"has_renders":84,"has_splat":84},[137],{"id":12,"title":138,"date":8,"status":139,"polarity":140,"category":141,"axes":142,"tags":144,"task_code":156,"related_runs":157,"delta_psnr":163,"delta_wallclock":164,"rank":165,"verdict":166,"impact_summary":167,"detail_path":168},"S2 forward 1 CB 化 — replay -12〜13% \u002F busy 69 → 80% (default flip)","stable","positive","experiment",[90,143],3,[145,13,146,147,148,149,150,151,152,153,154,155],"1x-roadmap-v2","forward-chained","command-buffer","hierarchical-scan","indirect-dispatch","radix-sort","gpu-busy","sync-tax","bit-exact","lego","unified-memory","2026-08 1× 工程表 v2 — AC 復帰後 計測 queue #2 (S2 gate)",[158,159,5,160,161,162],"s2-12500-noprof-r1","s2-12500-noprof-r2","s2gate-base-12500-r1","s2gate-base-12500-r2","s2gate-base-12500-r3","12.5k **同一 session 交互 A\u002FB** n=3 mean: S2 **36.167 dB** (36.154 \u002F 36.192 \u002F 36.155) vs default **36.160 dB** (36.107 \u002F 36.167 \u002F 36.205) = **+0.007 dB** (guard ±0.1 内、n=3 mean 判定 — n=1 判定なし)。S2 は prefix scan 部分こそ bit-exact だが end-to-end では既存の atomic 順序非決定性を継承するため、この guard は実質的な検査として機能している (default arm 自身の run 間振れ幅が ±0.05 dB ある)","replay bench (σ ≤1.5% の clean block、ABBA 交互 2 round): final regime **17.833 → 15.637 ms\u002Fiter = -12.3%**、iter5000 regime **16.600 → 14.395 = -13.3%** (有意閾値 max(2σ,1%) = 1.0〜1.8%)。頑健性確認の min-to-min 推定でも -10.6% \u002F -13.6%。12.5k full run **同一 session 交互 A\u002FB n=3 mean: 208.26 → 187.65s = -9.9%** (training elapsed では 205.10 → 184.61s = -10.0%)。参考値として counter session 当日 baseline 216.12s 比では -13.2%","high","s2-adopted-default-flip-chained-forward-and-hierarchical-scan","1× 工程表 v2 の並列性能軸で\u003Cstrong>本命に格上げ\u003C\u002Fstrong>されていた S2 を gate し、採用した。W2 が backward\u002Floss chain を 1 CB に畳んだのに対し S2 は forward 側を畳む — radix の per-pass CPU prefix scan を hierarchical GPU scan (bit-exact な整数 prefix sum) に、pair_count の host readback を indirect dispatch に置き換え、forward 全段を \u003Cstrong>1 CB \u002F 1 encoder \u002F 末尾 1 wait\u003C\u002Fstrong> に連結した。replay \u003Cstrong>final -12.3% \u002F iter5000 -13.3%\u003C\u002Fstrong> (閾値 1.0〜1.8% の 8〜13 倍)、12.5k full run は\u003Cstrong>同一 session 交互 A\u002FB (ABBAAB) n=3 で 208.26 → 187.65s = -9.9%\u003C\u002Fstrong>。\u003Cstrong>counter session が積み残した「busy 68.9%」は 79.8% まで回復\u003C\u002Fstrong>し、sync-tax は 5.21 → 2.86 ms\u002Fiter (-45%)。roadmap が注目トレードに挙げた「dispatch 26 → ~60 の overhead」は \u003Cstrong>busy ms\u002Fiter +0.20 = 回収額の 1\u002F11〜1\u002F24\u003C\u002Fstrong> と定量的にケリがつき、follow-up (radix reduce の histogram fold) は\u003Cstrong>総額 -1.1% 未満と判明したため S3 の後ろに defer\u003C\u002Fstrong>。\u003Cstrong>full run の Δ (-9.9%) が replay の Δ (-12.3%) を下回る一方、iteration 別に見ると -24.7% (iter1000) → -10.0% (iter12500) と単調に縮む\u003C\u002Fstrong> — S2 が削るのが\u003Cstrong>splat 数に依らない固定の per-iter host コスト\u003C\u002Fstrong>であることの直接証拠で、W1\u002FW2 (GPU 仕事量に比例する施策) とは pass-through の性質が異なる。msplat 同日等価 gap は \u003Cstrong>1.273× → 1.105×\u003C\u002Fstrong>。累計 (数学的同値の施策のみ) 766.07 → 187.65s = \u003Cstrong>-75.5%\u003C\u002Fstrong>。","\u002Ffindings\u002Fp2-s2-forward-chained\u002F",1790227325618]