erik/acdream
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

docs(runtime): close cost-budgeted streaming slice

Browse source 
Record the exact physical-display evidence, ownership fixes, budget semantics, and performance deltas that close Slice E. Reconcile the architecture, roadmap, milestones, issues, documentation map, and session instructions before the explicit F/G approval gate.

Co-authored-by: Erik Nilsson <erikn@users.noreply.github.com>
This commit is contained in:
Erik 2026-07-24 20:52:20 +02:00
parent 91e82c3c68
commit e7d9d6fa7d
12 changed files with 290 additions and 46 deletions
Download patch file Download diff file Expand all files Collapse all files

12
docs/plans/2026-04-11-roadmap.md
View file

@ -1604,8 +1604,8 @@ port in any phase — no separate listing here.
> above the existing specialized physical caches. Capped/uncapped nine-stop
> and dense-Arwic physical routes passed; third-visit cache residence was
> non-growing and no checkpoint retained staged or retiring bytes. [Residency
> report](../research/2026-07-24-slice-d-unified-residency-report.md). Slice E
> cost-budgeted streaming/retirement is active. E2 replaced the flat
> report](../research/2026-07-24-slice-d-unified-residency-report.md).
> **Slice E completed 2026-07-24:** E2 replaced the flat
> deferred list, count-only throttle, priority hunt, and unload bypass with one
> typed-meter scheduler over explicit stable destination/control/unload/Near/
> Far queues; production prices worker results before adoption and lifecycle
@ -1619,7 +1619,13 @@ port in any phase — no separate listing here.
> canonical reveal generation to an exact destination reservation, protects
> its share across all typed budget dimensions, removes forced incomplete
> materialization, and presents retail's centered five-second portal wait cue.
> E6 deterministic pressure and physical-local closeout is next. F/G and J
> E6 deterministic/resource suites, Release build, complete tests, capped and
> uncapped nine-stop routes, and pinned dense Arwic pass on exact binary
> `91e82c3c`. Every checkpoint converged without pending publication,
> retirement, upload, warmup, or class backlog. Capped/uncapped CPU p99 improved
> 34.7%/30.4% from Slice A and maximum frame allocation fell 84.7%/82.2%.
> [Streaming report](../research/2026-07-24-slice-e-cost-budgeted-streaming-report.md).
> F/G and J
> retain their explicit approval gates.
**Spec:** `docs/superpowers/specs/2026-07-05-modern-pipeline-design.md` (the

10
docs/plans/2026-05-12-milestones.md
View file

@ -93,7 +93,7 @@ The separately authorized modern-runtime performance program has completed
Slices A–D: corrected measurement, prepared-package bake/dedup, package-only
production streaming, and typed unified residency. Its physical capped,
uncapped, and dense connected gates pass without a visual-quality change.
Slice E0–E5 now define and enforce typed completion admission through explicit
Slice E is complete. E0–E5 define and enforce typed completion admission through explicit
stable priority queues, publish an immediate generation-scoped old-world
quiesce at login/portal boundaries, and advance full-window/recenter teardown
through exact metered retirement cursors. Render, physics, static, building,
@ -101,8 +101,12 @@ and EnvCell publication now resume from retained exact cursors under the same
frame budget. The reveal coordinator now owns the exact destination
generation/cell/radius reservation across every budget dimension; portal
transit never force-reveals incomplete content and surfaces retail's centered
wait cue after five seconds. E6 deterministic pressure and physical-local
closeout is next. Slices F/G remain behind their explicit ECS approval gate,
wait cue after five seconds. E6 deterministic/resource suites, Release build,
complete tests, capped/uncapped nine-stop routes, and pinned dense Arwic pass
on exact binary `91e82c3c`; all checkpoints converge with no stranded
publication, retirement, upload, warmup, or resource backlog. Capped/uncapped
CPU p99 is 34.7%/30.4% below Slice A and maximum frame allocation is
84.7%/82.2% lower. Slices F/G remain behind their explicit ECS approval gate,
and Slice J
remains behind its gameplay-owner approval gate.

30
docs/plans/2026-07-24-modern-runtime-architecture.md
View file

@ -2,7 +2,7 @@
**Date:** 2026-07-24
**Status:** Slices A–D complete. Slice E is active (E0–E5 complete; E6 gates next).
**Status:** Slices A–E complete. Slices F/G await their explicit approval gate.
**Scope:** Reconcile and sequence the existing Modern Pipeline (`MP`) and
Linux/headless (`LH`) tracks using the 2026-07-24 connected performance audit.
@ -724,9 +724,9 @@ checkpoint retained staged or retiring bytes. Evidence:
### Slice E — Cost-budgeted streaming and retirement
**Active plan:** [`2026-07-24-modern-runtime-slice-e-cost-budgeted-streaming.md`](2026-07-24-modern-runtime-slice-e-cost-budgeted-streaming.md).
**Completed plan:** [`2026-07-24-modern-runtime-slice-e-cost-budgeted-streaming.md`](2026-07-24-modern-runtime-slice-e-cost-budgeted-streaming.md).
**Checkpoint 2026-07-24:** E0 fixed the retail/adaptation contract and current
**Completed 2026-07-24:** E0 fixed the retail/adaptation contract and current
owner inventory. E1 landed the validated work profile, deterministic meter and
completion charge model. E2 cut execution over to stable priority FIFOs. E3
added immediate generation quiescence and cursored retirement. E4 advances
@ -734,8 +734,17 @@ render/physics/static publication from retained exact receipts. E5 replaced
mutable priority-radius state with the reveal coordinator's exact
generation-scoped destination reservation, protected its share across every
typed budget dimension, removed forced early materialization, and added
retail's centered five-second portal wait cue. E6 deterministic pressure and
physical-local closeout is next.
retail's centered five-second portal wait cue.
E6's deterministic suites, Release build, complete tests, capped/uncapped
nine-stop routes, and pinned dense-Arwic route pass. The connected gate
corrected stale cache-hit staging and separated quiesced destination spatial
residency from gameplay visibility. Every canonical checkpoint converged with
zero pending publication/retirement/backlog, staged upload, composite warmup,
or retiring resource. Against Slice A, capped/uncapped CPU p99 improved
34.7%/30.4% and largest frame allocation fell 84.7%/82.2%; against Slice C,
CPU p99 improved 21.4%/16.1% and allocation fell 33.5%/9.7%. Evidence:
[`../research/2026-07-24-slice-e-cost-budgeted-streaming-report.md`](../research/2026-07-24-slice-e-cost-budgeted-streaming-report.md).
**Purpose:** Remove update-thread portal transactions.
@ -748,11 +757,12 @@ physical-local closeout is next.
- Preserve FIFO within priority/generation and exact retry receipts.
- Connect reveal to the single destination-generation readiness state.
**Gate (rescoped, 2026-07-24 review):** zero `viewport-before-ready`; no
single-frame publication/retirement transaction exceeds the configured
time/byte/entity budgets; portal-window frame p99 and single-frame
allocation materially improve versus the post-Slice-A baseline; no stranded
old generation, staged upload, collision, effect, or GPU owner. The
**Gate (rescoped, 2026-07-24 review; completed):** zero
`viewport-before-ready`; no scheduler continuation after a configured
time/byte/entity limit (the documented first-indivisible-operation rule is
named explicitly); portal-window frame p99 and single-frame allocation
materially improve versus the post-Slice-A baseline; no stranded old
generation, staged upload, collision, effect, or GPU owner. The
program-level 16.67 ms p99 / 33.3 ms maximum targets are §11 gates verified
at Slice H — Slice E cannot own them while the audit's own P2 costs
(whole-world partition, unconditional UI/diagnostic work) are assigned to

39
docs/plans/2026-07-24-modern-runtime-slice-e-cost-budgeted-streaming.md
View file

@ -1,6 +1,6 @@
# Modern Runtime Slice E — Cost-budgeted streaming and retirement
**Status:** active — E0/E1/E2/E3/E4/E5 complete; E6 gates and closeout next
**Status:** complete — E0/E1/E2/E3/E4/E5/E6 landed and gated 2026-07-24
**Program:** `docs/plans/2026-07-24-modern-runtime-architecture.md`
**Baseline:** Slice D closeout commit `66690805`
**Behavior contract:** no visual-quality or retail-behavior reduction
@ -398,14 +398,45 @@ E6, not E5.
- Update architecture, inventory, issues, roadmap, divergence register, and
durable memory.
**Complete 2026-07-24.** Deterministic budget, pressure, cancellation,
stale-generation, retry, callback-reentrancy, GUID-reuse, quiescence, and
resource-ownership gates pass. The Release solution builds and the complete
suite passes 8,164 tests with 5 intentional skips.
The exact `91e82c3c6850fcedbf20322b2ecf4fdcd11a2b2a` Release binary passed
capped and uncapped nine-stop routes plus pinned dense Arwic on the physical
local AMD display. Every destination materialized only after the canonical
render/composite/collision predicate, every client exited gracefully, and
every canonical checkpoint ended with zero pending publication, retirement,
worker/destination/class backlog, staged upload, composite warmup, residency
staging/requested-GPU, or retiring bytes.
The gate found and corrected two connected-only ownership defects. CPU
mesh-cache hits now stage only for a live exact owner, preventing evicted work
from recreating stale uploads. Loaded spatial residency is now distinct from
world availability, allowing quiesced destination live objects to prepare
their render projection without becoming drawable, collidable, pickable,
targetable, audible, or simulated before reveal.
Relative to Slice A, capped/uncapped CPU p99 improved 34.7%/30.4% and largest
frame allocation fell 84.7%/82.2%. Relative to Slice C, CPU p99 improved
21.4%/16.1% and largest frame allocation fell 33.5%/9.7%. Full evidence:
[`../research/2026-07-24-slice-e-cost-budgeted-streaming-report.md`](../research/2026-07-24-slice-e-cost-budgeted-streaming-report.md).
## 10. Acceptance
Slice E closes only when:
1. No priority or unload path bypasses all work budgets.
2. No frame exceeds any configured count/byte/entity dimension.
3. Every time overrun names one indivisible operation; none remain above the
configured time ceiling at closeout.
2. No frame admits another operation after a configured count/byte/entity
dimension is exhausted. The documented first-indivisible-operation rule is
the sole oversize path and remains named.
3. Every wall-time overrun names one indivisible operation and the scheduler
admits no later operation after its deadline. Because the meter uses elapsed
wall time on a non-real-time OS, thread preemption can make an otherwise
bounded atomic operation observe more than 2 ms; the physical gate therefore
also requires materially improved route/portal p99 and no repeatable
algorithmic tail.
4. Old generations become unavailable, tick-frozen, and audio-silent at the
hard transition edge.
5. Retirement/publication retries resume exact cursors without replay.