Fr0zkaandClaude Opus 5.5 c152708a67 wip(standalone): GPU step 15 -- cold-start drain: hybrid whole view 3.2 -> 1.25-1.68 s ...

Timeline showed the post-export window was frame-paced: the 4-batch / 2 ms
per-frame upload budget held CPU-fallback tiles up to 755 ms after they
finished, and 35 GPU batches drained one per frame. Before the first whole
view, GPU work now drains as one batch and CPU fallback starts with GPU input
preparation (--gpu-far-no-cold-drain = paced control).

Luna's race: CPU 1.393 s vs GPU 1.677 s. My interleaved runs (Step15Timeline,
RunGpuFarViewer): GPU 2.902 (first launch) / 1.252 / 1.209 s; CPU 2.196 /
2.250 / 2.061 s. The CPU viewer still uses the paced upload budget at cold
start (stream 674 -> 2250 ms in my runs vs 741 -> 1422 in Luna's), so the
race is not yet fair. Audits: hashes, walk 5,642/5,642, 0 diffs vs CPU/WARP.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MmSoF64BbRFmk9C3p1VsBJ

2026-10-07 08:47:36 +02:00

..

final

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

gpu_field_port

wip(standalone): GPU step 15 -- cold-start drain: hybrid whole view 3.2 -> 1.25-1.68 s

2026-10-07 08:47:36 +02:00

gpu_first_port

wip(standalone): GPU step 14 -- warm-up 2.04 -> 1.05 s; cold path now waits on streaming, not shaders

2026-10-07 08:03:00 +02:00

round2

standalone: shape-world look round 2 -- shards gone, calmer layered relief, still 0.72 s / 1.6 s cold

2026-10-04 04:09:59 +02:00

speed_round

wip(standalone): GPU step 6 -- near field on GPU, exact; cold start still slower than CPU

2026-10-06 20:03:10 +02:00

chamber_shape_final.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

chamber_shape_floorclip.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

chamber_shape_sdf.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

chamber_shape.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

farclip_shape_final.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

farclip_shape.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

passage_shape_final.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

passage_shape_floorclip.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

passage_shape_sdf.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

passage_shape.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

REPORT.md

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

tunnel_shape_final.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00

tunnel_shape.png

research(standalone): shape world -- 64 m full detail in 0.96 s, whole view in 2.27 s, cold; look not there

2026-10-03 11:44:45 +02:00