perf(content): close prepared asset cutover
This commit is contained in:
Erik
parent
b1ad4b7c0a
commit
a564c4b782
11 changed files with 428 additions and 66 deletions
|
|
@ -1562,9 +1562,10 @@ port in any phase — no separate listing here.
|
|||
> remains DEFERRED**: ECS is the *throughput* lever (raise avg FPS ~230→300), a
|
||||
> DIFFERENT goal from steadiness — revisit ONLY if the user wants a higher raw FPS
|
||||
> number and accepts the rewrite risk. Rust rejected. Parked (resume from the table
|
||||
> below): MP1b pak/bake is now complete and the corrected physical-local
|
||||
> baselines have landed, so MP1c is active. The remaining parked items are
|
||||
> MP-Alloc hard sites and MP1c. MP2 has shipped. Full rationale:
|
||||
> below): MP1b pak/bake and MP1c production cutover are now complete. The
|
||||
> remaining parked items are MP-Alloc hard sites; typed residency and
|
||||
> cost-budgeted streaming continue as Slices D/E in the modern-runtime plan.
|
||||
> MP2 has shipped. Full rationale:
|
||||
> `project_mp_track_findings.md`.
|
||||
> The 2026-07-18 portal/resource-lifetime repair does not resume MP's ECS, pak,
|
||||
> or throughput redesign. It is corrective ownership/reclamation work for a
|
||||
|
|
@ -1588,10 +1589,15 @@ port in any phase — no separate listing here.
|
|||
> portal windows, observer-effect measurement, binary provenance, and
|
||||
> contention stacks. The physical runs confirm fast steady rendering but also
|
||||
> preserve repeated-portal memory plateau failures, ~200 MiB allocation frames,
|
||||
> and thousands of invalid typed-Setup probes as MP1c's before-state. Slice C
|
||||
> is unblocked and active. Slice B EnvCell dedup is complete; its
|
||||
> and thousands of invalid typed-Setup probes as MP1c's before-state. Slice B
|
||||
> EnvCell dedup is complete; its
|
||||
> [full-bake report](../research/2026-07-24-slice-b-full-bake-report.md)
|
||||
> records the 42.6× result and corrected geometry-identity collision.
|
||||
> **Slice C completed 2026-07-24:** production streaming consumes the validated
|
||||
> bake-tool-3 package, all physical connected and user-visual gates passed,
|
||||
> process allocation fell 41.8–61.9%, GC pause time fell 48.4–54.1%, and
|
||||
> invalid typed-Setup exceptions fell to zero. [Cutover
|
||||
> report](../research/2026-07-24-slice-c-prepared-asset-cutover-report.md).
|
||||
|
||||
**Spec:** `docs/superpowers/specs/2026-07-05-modern-pipeline-design.md` (the
|
||||
umbrella design — read it first). **Goal:** smoothness first (no frame over
|
||||
|
|
@ -1616,7 +1622,7 @@ hitch and follows.
|
|||
| MP1b | Pak format + `acdream-bake` CLI + mmap `PakReader` + equivalence/full-scale gate | ✅ **SHIPPED 2026-07-24** — 729,888 EnvCell keys → 17,117 unique geometries + 712,771 aliases (42.6×), 751,141 total keys / 38,370 physical blobs, zero failures, 28,192.4 MiB, 81.4 s validated atomic publish; full-DAT gate also corrected a real collision in WB's legacy geometry hash. [Report](../research/2026-07-24-slice-b-full-bake-report.md) |
|
||||
| **MP-Alloc (safe batch)** | Reuse per-frame buffers: anim pose, particle draw-list, interior partition, animatedIds/drawableCells sets | ✅ SHIPPED + USER-GATED 2026-07-05 — dense-town frame-time spikes 20–87ms → 6–10ms, alloc spikes (30–75MB single-frame) eliminated, gen2 GC 5–11/window → ~0; user confirms FPS steady. 4 commits `b8c05e2b`..`91afea24` |
|
||||
| MP-Alloc (hard sites) | EnvCell settled-camera rebuild gate + physics `Transition` pooling — the residual steady ~1.6MB/frame | ⚪ OPTIONAL follow-up — would lower the median too, but the wild-swing complaint is already resolved; each needs its own careful gate (batch correctness / physics faithfulness) |
|
||||
| MP1c | Streaming cutover to pak + hitch gate | **ACTIVE 2026-07-24** — physical-local Slice A gate committed; baseline: uncapped CPU p50/p95/p99 1.520/5.694/7.825 ms, GPU 0.897/3.239/3.406 ms, portal allocation max 212.7 MiB, repeated-Caul working-set growth 295.1 MiB. Slice B package foundation is complete |
|
||||
| MP1c | Streaming cutover to pak + hitch gate | ✅ **SHIPPED 2026-07-24** — typed package-only production source, explicit Setup probes, non-allocating Portal-first lookup, exact translucency metadata, and ordered teardown; capped/uncapped/dense physical routes + user visual gate pass. Uncapped CPU p99 7.825→6.496 ms, GPU p99 3.406→2.706 ms, portal frame max 202.9→39.9 MiB, process allocation −41.8%, zero invalid Setup probes. [Report](../research/2026-07-24-slice-c-prepared-asset-cutover-report.md) |
|
||||
| MP2 | Retail particle distance/cell-view degradation: DAT-authored range, exact infinite/finite off-view update branches, emitter-first render scan, user-requested 2× default range | ✅ SHIPPED 2026-07-17 — Aerlinthe hotspot `0x32000223` resolves to retail 64m and defaults to 128m; landscape/entity streaming distance unchanged |
|
||||
| MP3 | Arch ECS render world + delta submission (the 300-FPS lever) | ⚪ — note: does NOT fix the steady-state GC churn (that's MP-Alloc); MP3 is the draw-submission throughput lever |
|
||||
| MP4 | Zero-alloc frame loop + flat physics data (residual, post-MP-Alloc) | ⚪ hard-queued behind M1.5 #137 |
|
||||
|
|
|
|||
|
|
@ -2,8 +2,7 @@
|
|||
|
||||
**Date:** 2026-07-24
|
||||
|
||||
**Status:** Slice A complete on the physical reference display; Slice B
|
||||
complete. Slice C is unblocked and active.
|
||||
**Status:** Slices A–C complete. Slice D is next and not yet started.
|
||||
|
||||
**Scope:** Reconcile and sequence the existing Modern Pipeline (`MP`) and
|
||||
Linux/headless (`LH`) tracks using the 2026-07-24 connected performance audit.
|
||||
|
|
@ -631,6 +630,15 @@ collision in WorldBuilder's legacy 31× geometry hash. Evidence:
|
|||
single-frame allocation and p99 materially improve from the 2026-07-24 baseline;
|
||||
connected nine-stop route and screenshots pass.
|
||||
|
||||
**Completed 2026-07-24:** production streaming now consumes the validated
|
||||
format-1/bake-tool-3 package through `IPreparedAssetSource`, with no implicit
|
||||
live-DAT mesh fallback. Capped, uncapped, and dense-Arwic physical routes,
|
||||
installed-DAT equivalence, lifecycle assertions, graceful teardown, the full
|
||||
Release suite, and the user's visual gate passed. Process allocation fell
|
||||
41.8–61.9%, GC pause time fell 48.4–54.1%, uncapped CPU/GPU p99 improved
|
||||
17.0%/20.6%, and invalid typed-Setup exceptions fell to zero. Evidence:
|
||||
[`../research/2026-07-24-slice-c-prepared-asset-cutover-report.md`](../research/2026-07-24-slice-c-prepared-asset-cutover-report.md).
|
||||
|
||||
**Amendment (2026-07-24 review).** Binding contracts for this slice:
|
||||
|
||||
- **Precedence contract.** The LINQ being replaced is load-bearing content
|
||||
|
|
@ -1129,13 +1137,14 @@ an improvement" carries the same evidence burden as a regression claim.
|
|||
|
||||
## 14. Current execution point
|
||||
|
||||
Slice A's tooling and authoritative physical-local evidence are complete, and
|
||||
Slice B's prepared-content bake is complete. Slice C is active. Its comparison
|
||||
baseline is the committed physical-local capture at
|
||||
`49f3a48ee63b82d00731b0f91497930925279c86`: uncapped CPU p50/p95/p99
|
||||
1.520/5.694/7.825 ms, GPU p50/p95/p99 0.897/3.239/3.406 ms, portal allocation
|
||||
max 212.7 MiB, process allocation mean 50.0 MB/s, and a repeated-Caul working
|
||||
set plateau failure of 295.1 MiB.
|
||||
Slices A–C are complete. Slice C replaced production live-DAT world-mesh
|
||||
extraction with the validated prepared source and passed all connected,
|
||||
performance, lifetime, automated, and user-visual gates. Its physical
|
||||
uncapped result is CPU p50/p95/p99 1.409/4.909/6.496 ms, GPU
|
||||
0.832/2.510/2.706 ms, a 39.9 MiB largest frame allocation, zero invalid Setup
|
||||
probe exceptions, and a passing warm repeated-Caul memory assertion. Slice D
|
||||
is the next execution point; it unifies accounting and policy around the
|
||||
specialized caches that remain, rather than replacing their working storage.
|
||||
|
||||
The intended order is therefore:
|
||||
|
||||
|
|
|
|||
|
|
@ -2,15 +2,18 @@
|
|||
|
||||
**Date:** 2026-07-24
|
||||
|
||||
**Status:** active — C1 and C2 landed; C3 implemented, C4 connected gate next
|
||||
**Status:** complete — automated, connected, performance, and user visual
|
||||
gates passed 2026-07-24
|
||||
|
||||
**Checkpoint ledger:**
|
||||
|
||||
- C1 typed prepared-source boundary — `c42f93b3`
|
||||
- C2 production renderer injection — `f05afc07`
|
||||
- C3 activation probes and lookup precedence — `230a7df4`; automated gate
|
||||
green, connected exception-count gate remains in C4
|
||||
- C4 connected equivalence and performance closeout — pending
|
||||
green
|
||||
- C4 prepared-source diagnostics — `b1ad4b7c`
|
||||
- C4 connected equivalence and performance closeout —
|
||||
[`../research/2026-07-24-slice-c-prepared-asset-cutover-report.md`](../research/2026-07-24-slice-c-prepared-asset-cutover-report.md)
|
||||
|
||||
**Parent:** [`2026-07-24-modern-runtime-architecture.md`](2026-07-24-modern-runtime-architecture.md),
|
||||
Slice C / MP1c
|
||||
|
|
@ -300,6 +303,32 @@ If pixels or behavior differ, the slice is not complete even when performance
|
|||
improves. If timing improves but heap reservation still fails, keep the failure
|
||||
open and attribute it before considering Slice C closed.
|
||||
|
||||
### C4 result — completed 2026-07-24
|
||||
|
||||
The format-1/bake-tool-3 production package validated with 751,141 keys and
|
||||
the exact installed DAT iterations. Installed-DAT seam/equivalence tests, all
|
||||
three connected physical-display routes, canonical lifetime assertions, and
|
||||
the complete Release suite passed. The user accepted the prepared path's
|
||||
geometry, materials, textures, and world presentation.
|
||||
|
||||
Against the committed Slice A physical baseline:
|
||||
|
||||
- capped CPU p95/p99 improved 15.8%/16.9%, GPU p95/p99 17.4%/20.0%;
|
||||
- uncapped CPU p95/p99 improved 13.8%/17.0%, GPU p95/p99 22.5%/20.6%;
|
||||
- dense-Arwic CPU/GPU p99 improved 12.0%/16.8%;
|
||||
- process allocation fell 54.9% capped, 41.8% uncapped, and 61.9% dense;
|
||||
- GC pause time fell 52.8%, 48.4%, and 54.1%;
|
||||
- the invalid typed-Setup exception count fell to zero in every route;
|
||||
- the largest frame-thread allocation fell from roughly 203 MiB to 46.8 MiB
|
||||
capped and 39.9 MiB uncapped;
|
||||
- both warm repeated-Caul memory comparisons passed the unchanged assertion.
|
||||
|
||||
Windows reports clean memory-mapped package pages in working set, so total
|
||||
working set is higher even though private memory and ownership remain bounded.
|
||||
That is file cache, not retained world state. Exact measurements and the
|
||||
artifact policy are recorded in the
|
||||
[Slice C report](../research/2026-07-24-slice-c-prepared-asset-cutover-report.md).
|
||||
|
||||
## 4. Review checklist per checkpoint
|
||||
|
||||
Because this task is running without delegated reviewers, the primary engineer
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue