[{"data":1,"prerenderedAt":164},["ShallowReactive",2],{"run:c1-12500-off-prof":3,"run-findings:c1-12500-off-prof":135},{"run":4,"config":17,"metrics":80,"curve":86,"assets":134},{"id":5,"title":5,"subtitle":6,"eyebrow":7,"date":8,"workspace":9,"commit":10,"tags":11,"verdict":16},"c1-12500-off-prof","M1 C1: per-kernel gpu_ms 再計測 (submission=per_kernel \u002F forward_submission=per_stage \u002F pipelining=off、post-S2\u002FS3 kernel)","Run summary · 2026-09-seminar05-milestones M1 (C1 per-kernel 再計測 + C2 残差 bubble 帰属)","2026-08-30","splat","705e3f8",[12,13,14,15],"p3-kernel-share-post-s3","m1","gpu-gate","lego","partial",[18,21,24,27,30,33,36,39,42,45,48,51,54,57,60,63,66,69,72,75,77],{"key":19,"value":20},"dataset","\u002FUsers\u002Fotkrickey\u002Fdev\u002F3dgs-workspace\u002Fdatasets\u002Fnerf_synthetic\u002Flego",{"key":22,"value":23},"gt_convention","Premultiplied",{"key":25,"value":26},"iterations","12,500",{"key":28,"value":29},"seed","42",{"key":31,"value":32},"capacity","1,000,000 splats",{"key":34,"value":35},"sh_degree","3",{"key":37,"value":38},"loss","L1Ssim",{"key":40,"value":41},"lambda","0.200",{"key":43,"value":44},"ssim","window=7 sigma=1",{"key":46,"value":47},"backend.backward","Simd",{"key":49,"value":50},"backend.loss_path","Gpu",{"key":52,"value":53},"backend.emit_pairs","Original",{"key":55,"value":56},"backend.forward","F32",{"key":58,"value":59},"backend.dispatch","PerStage",{"key":61,"value":62},"backend.radix_scan","GpuHierarchical",{"key":64,"value":65},"backend.binning","Resident",{"key":67,"value":68},"backend.tile_sort","GlobalRadix",{"key":70,"value":71},"backend.submission","PerKernel",{"key":73,"value":74},"backend.target_cache","true",{"key":76,"value":74},"backend.shader_fast_math",{"key":78,"value":79},"os_build","26.6.2 (25G83)",{"psnr":81,"wallclock":82,"wallclock_regress":83,"splats":84,"final_loss":85},36.08714294433594,"3m 25s",false,408006,"5.521357e-3",{"loss":87,"splats":118},{"iters":88,"values":103},[89,90,91,92,93,94,95,96,97,98,99,100,101,102],1,1000,2000,3000,4000,5000,6000,7000,8000,9000,10000,11000,12000,12500,[104,105,106,107,108,109,110,111,112,113,114,115,116,117],0.5938452482223511,0.05445867031812668,0.014137223362922668,0.008451830595731735,0.007436282001435757,0.006633849814534187,0.006221305578947067,0.005989145487546921,0.005858478136360645,0.00572655675932765,0.005663991905748844,0.005585349164903164,0.005523400381207466,0.0055213565938174725,{"iters":119,"values":120},[89,90,91,92,93,94,95,96,97,98,99,100,101,102],[121,122,123,124,125,126,127,128,129,130,131,132,133,84],5207,8243,181374,297067,349683,375082,389255,397002,401504,403649,404755,405838,407140,{"has_renders":83,"has_splat":83},[136],{"id":12,"title":137,"date":8,"status":138,"polarity":139,"category":140,"axes":141,"tags":143,"task_code":155,"related_runs":156,"delta_psnr":158,"delta_wallclock":159,"rank":160,"verdict":161,"impact_summary":162,"detail_path":163},"M1\u002Fv2 — post-S2\u002FS3 per-kernel share 再計測: backward_raster 33.9%→40.1% (M4\u002FA7 後)、radix 26.4%→12.0% (M4 の絶対削減を反映) → M3 GO 判定は不変。C2 残差 4.54s は driver latency \u002F CB gap に分解不能 (計装欠如、file:line 提案のみ)","stable","mixed","measurement",[89,142],3,[13,144,145,146,147,148,149,150,151,15,152,153,154],"m7","seminar05-milestones","kernel-share","per-kernel-gpu-ms","backward-raster","m3-go-gate","residual-bubble","gputime","post-s2-s3","v2","post-m4-a7","2026-09-seminar05-milestones M1 (C1 per-kernel 再計測 + C2 残差 bubble 帰属)",[5,157],"c1v2-12500-off-prof","n=1: 36.087 dB (self-eval val mean, 100 views)。p2-s3 の off\u002Fdepth2 n=3 mean (36.185 \u002F 36.202 dB) の guard 内。本 run は submission=per_kernel\u002Fforward_submission=per_stage という別 regime なので PSNR の直接比較は非対象 (kernel は bit-exact のため理論上は同水準になるはず、というだけの参考値)","train elapsed 205.78s (n=1)。**production baseline (depth2 default, 151.70s) や S3 off baseline (185.21s) と直接比較しない** — per_kernel\u002Fper_stage submission は CB 数が既定の 15〜27\u002Fiter → 15 CB\u002Fiter 超 (最終 15,001 cb from iter1 window 相当) に増え、busy% が 68.9% まで下がる**測定専用 regime**であり、本 run の wallclock 自体に意味はない。意味があるのは kernel 単体の gpu_ms とその share のみ","high","m3-go-backward-raster-33.9pct \u002F c2-instrumentation-gap-proposal-only","M3 (per-splat backward, brush PR #401 型) の GO\u002FNO-GO 閾値「backward_raster share ≥ 25%」を判定した。post-S2\u002FS3 kernel (hierarchical radix scan \u002F resident binning) を submission=per_kernel\u002Fforward_submission=per_stage\u002Fpipelining=off (per-kernel breakdown が測れる唯一の regime) で 1 run 実測した結果 \u003Cstrong>backward_raster (+project_back 統合) share 33.86%\u003C\u002Fstrong> で \u003Cstrong>閾値を大きく超過 → M3 は GO\u003C\u002Fstrong>。\u003Cstrong>share × 0.25 (brush の -25% 主張をそのまま適用したナイーブな期待値) = 8.47%\u003C\u002Fstrong>。ただしこれは本 run の GPU busy (68.9%、production の 91.3% とは別 regime) に対する share であり、production (depth2 default) の wallclock に対する期待 Δ は \u003Cstrong>33.86% × 91.3% × 0.25 ≈ 7.7%\u003C\u002Fstrong> とみるほうが妥当 (kernel 自体の gpu_ms は submission wrapping にほぼ不変という counter-session の事後検証はあるが、n=1 点推定であることと合わせて上限値として扱う)。\u003Cstrong>pre-S2\u002FS3 (counter-session) との比較では share 構成はほぼ不変\u003C\u002Fstrong> (backward_raster 33.6%→33.86%、radix 26.3%→26.40%、rasterize 11.9%→11.64%、adam 13.8%→12.68%、ssim 10.0%→10.95%) — S2 のhierarchical scan は「radix の CB\u002Fwait 数」を畳んだだけで kernel 自体の GPU 時間配分は変えていない、という roadmap の当初の想定を実測で裏づけた。C2 は M0 の並行会計訂正 (encode(B) 4.66→3.41s、残差 3.28→4.54s、nested label 二重計上の是正) を踏まえ、\u003Cstrong>残差 4.54s (depth2 sync-tax の 32%) を driver 投入 latency と CB 間 gap に分解しようとしたが、既存計装 (`gputime.rs`) は `kernelStartTime`\u002F`kernelEndTime` を objc bind しているのみで収集しておらず、per-CB の生 interval 列も外部に露出していない\u003C\u002Fstrong> ため\u003Cstrong>分解不能\u003C\u002Fstrong>と判明した。file:line 付きの追加案を §C2 に残す (実装はしていない、GPU run はこれ以上起動していない)。","\u002Ffindings\u002Fp3-kernel-share-post-s3\u002F",1790227323336]