release: add versioned binary manifest (W5)

.agents/NOW.md

@@ -26,11 +26,11 @@ Work: exact-chunks on main `1ce0d662b`; sm_120 measured at `3d2581551`.
 | CPU levers (`QUANT-GGUF-CIQ-GEMM`) | Profile DONE: decode **47% threadpool sync**, prefill **~39% paged attn**. **G5 not next** | Parakeet encoder; attn dtype hoist |
 | Supported-models list | **LANDED**: FEATURES arch table CI-bound (33 archs) | — |
 | `/v1/videos` OpenAI shape | **MERGED** (#71): Sora `model`/`size`/`seconds` + `GET /{id}/content` | `row/SERVE-VIDEOS-REFS` PR open: reference conditioning |
-| Vulkan 27B decode | **4.285 vs 4.35**. Host ~3 ms/tok not 5.3; pipelined submit -1.41 ms/tok 5/5; spin REJECTED | NOT met. Next: 900 barriers/tok |
+| Vulkan 27B decode | **4.285 vs 4.35**; host ~3 ms/tok, pipeline -1.41 ms (5/5), spin rejected | NOT met: 900 barriers/tok |
 | `BACKEND-ROCM` | **(b) fix in; #140 gfx1201 hipBLAS + Gemma-4 MoE landed (contributor, authorship-preserved); W0 green 4 archs** | compile + M2 ([spec](specs/rocm-unified-memory-b.md)) |
 | TP spike #287 (PR #143) | **TP-W1 LANDED**: rank-group table + TP handle (6/6); DSR leak FIXED (unblocks #127/#154/#155) | TP-W2 (linears + loader) |
-| Release | SPIKE; 30/30 | #129 |
-| Surface coverage (`ARCH-ONE-SURFACE`) | ROW 8 + #139 IN; **ROW 6 LANDED (#137): embeddings LIVE — `LlamaModel` arch, PoolingRunner in the step, `vllm_embed` v15, `/v1/embeddings`, fold gate 4/4-231, 9 kills** | Real-checkpoint oracle cosine residual |
+| Release | **ACTIVE; W5 19/19+10/10; contract 30/30** | #141; artifacts pending |
+| Surface coverage (`ARCH-ONE-SURFACE`) | ROW 8 + #139; **embeddings live (#137): model, runner, ABI v15, endpoint, fold 4/4-231** | Real-checkpoint oracle cosine |
 
 In-flight, default-OFF, not pushed: see state.md.
 

.agents/benchmark-record.md

@@ -16758,6 +16758,19 @@ MEASURE 243-248 GB/s -- above 230.3 -- so 230.3 is a ceiling for THAT SHAPE
 ceiling sits ~7% below the same kernel's on layer weights is UNEXPLAINED, and is
 the same lone buffer the 20x bimodal collapse attaches to.
 
+## Rolled out of the status page on 2026-08-09
+
+Moved verbatim from `docs/STATUS.md`; the keyed capability row remains the
+public status surface.
+
+## 2026-08-08 — Gemma4 ROCm fused helpers via portable vt:: seam (#154)
+
+Model files (`gemma4.cpp`, `gemma4_moe.cpp`) no longer call `vt::rocm::*` directly.
+Fused paths go through `include/vt/fused_ops.h` (`vt::RmsNormPlusAdd`,
+`DualRmsNormPlusRes`, `GeluMulSeparate`, `MatmulBTAlphaBeta`, `MatmulBTFp8Channel`,
+`ExpertGeGLUBf16TopKM1`). ROCm fast path under `VLLM_CPP_HIP`; non-HIP stubs for
+peer/pin/resident upload. `check-device-leakage` holds baseline.
+
 ## Laguna-S-2.1 MoE STATUS cell, relocated verbatim (2026-08-09)
 
 MOVED, NOT EDITED. `docs/STATUS.md` carried this as a single 33,211-char table

.agents/coordination.md

@@ -118,23 +118,19 @@ without the selected contention proof for their entire run are discarded.
 
 ## Active claims
 
-**Server binary release-matrix spike (`ENG-RELEASE-BINARIES`, 2026-08-07,
-`CLAIM-ENG-RELEASE-BINARIES-SPIKE`).** Codex (GPT-5), isolated worktree
+**Server binary release manifest W5 (`ENG-RELEASE-BINARIES`, 2026-08-08,
+`CLAIM-ENG-RELEASE-BINARIES-W5`).** Codex (GPT-5), isolated worktree
 `/home/mudler/_git/vllm.cpp-release-binaries`, branch
-`row/ENG-RELEASE-BINARIES`, base `origin/main` `9cb799cb`. Records/spec-only
-claim for issue #117 and the user-approved hybrid stable/preview downloadable
-server-bundle matrix, including the fully-static feasibility boundary. Owns
-only NEW `.agents/specs/release-binary-matrix.md`, the new
-`ENG-RELEASE-BINARIES` engine-matrix row, its single roadmap portfolio link,
-this claim, the required current-state checkpoint in `docs/STATUS.md` and
-`docs/BENCHMARKS.md`, the engine inventory count pin in
-`scripts/check-agent-record.py`, `.agents/NOW.md`, and the append-only
-`.agents/state.md`.
-No release workflow, CMake/package implementation, source, test, model, kernel,
-generated fixture, download, build, GPU, or service change is in scope. Draft
-PR [#129](https://github.com/mudler/vllm.cpp/pull/129) is the claim. The spike
-contract incorporates the user-reviewed fat-CUDA/adaptive-CPU primary-artifact
-revision and awaits fresh review; the row stays `SPIKE`.
+`row/ENG-RELEASE-BINARIES`, base `origin/main` `b38f78a7`. PR #129 merged the
+accepted release matrix; the current claim is W5 only: a versioned manifest
+schema, deterministic generator/validator, canonical fixtures, mutation tests,
+and the narrow release-contract/record surfaces those tests require. No W1-W4
+CUDA/CPU implementation, W6 archive/install target, release workflow,
+published artifact, download, GPU, model, kernel, or service change is in
+scope. Draft PR #141 is the claim. W5 is implemented and green at 19/19 after
+the fresh review's ten production-removal mutations were killed; the row is
+`ACTIVE`, while every archive, staged smoke, runtime, correctness, performance,
+install/package and publication claim remains pending.
 
 **CPU grouped keep-quant GEMM activation-dtype P0 (`QUANT-GGUF-CIQ-GEMM`,
 2026-08-06, `CLAIM-QUANT-GGUF-CIQ-GROUPED-DTYPE`).** Claude Code
@@ -1413,7 +1409,7 @@ table, tests, CMake. Details in the state-log entry of the same date.
 
 | Claim | Row IDs | Agent | Worktree / remote dir | Branch | Owned scope | State | Last update |
 |---|---|---|---|---|---|---|---|
-| `CLAIM-ENG-RELEASE-BINARIES-SPIKE` | `ENG-RELEASE-BINARIES` | Codex (GPT-5) | `/home/mudler/_git/vllm.cpp-release-binaries`; records/spec only, no build/GPU/download/service work | `row/ENG-RELEASE-BINARIES`, base `9cb799cb`, draft PR #129 | Accepted release-bundle spike plus its matrix/roadmap/coordination/state/NOW and public-doc lifecycle checkpoint; no CMake, workflow, source, test, or artifact implementation | `ACTIVE` | 2026-08-07 — user-reviewed revision complete: primary fat CUDA + adaptive CPU per host ABI, optional per-SM diagnostics; row stays `SPIKE`; awaiting fresh review |
+| `CLAIM-ENG-RELEASE-BINARIES-W5` | `ENG-RELEASE-BINARIES` (`ACTIVE`; W5 only) | Codex (GPT-5) | `/home/mudler/_git/vllm.cpp-release-binaries`; CPU/build-time metadata tooling only, no GPU/download/service work | `row/ENG-RELEASE-BINARIES`, base `b38f78a7`, claim commit `29107d0b`, draft PR #141 | W5 implemented: versioned manifest schema + deterministic generator/validator, canonical synthetic CPU/CUDA fixtures, fail-closed mutation tests, release checker/registration and required record/doc checkpoints. Excludes W1-W4, W6-W13, archives, install/package/publish workflows and runtime artifacts | `ACTIVE` | 2026-08-08 — W5 19/19; fresh-review production removals 10/10 killed; accepted release suite 30/30; no archive or real runtime/correctness/performance evidence |
 | `CLAIM-KIMI-LINEAR-W0` | `MODEL-TEXT-kimi-linear-kimi-linear-for-causal-lm` (stays SPIKE — dedicated W0 spike) | Claude Code (opus-4-8) | isolated worktree `.claude/worktrees/agent-a771cc029e6843be0`; CPU-only, records-only — NO build, NO GPU, NO download (two GPU jobs queued ahead; the W0 GPU golden capture is a SEPARATE later step) | branch `worktree-agent-a771cc029e6843be0`, base `origin/main` HEAD `10dd23ee` | The FULL dedicated W0 spike for `KimiLinearForCausalLM` (Kimi-Linear-48B-A3B) per the spike-first protocol, so W1 implementation can start immediately. Owns ONLY: NEW `.agents/specs/kimi-linear.md`, the `MODEL-TEXT-kimi-linear-*` matrix row (Spike link → the dedicated spec + note; STAYS `SPIKE` because it is actively claimed — protocol forbids a claimed row from being `READY`; checklist mark stays `📋`), this claim row + the narrative block below, the roadmap breadth note, `docs/STATUS.md`/`docs/BENCHMARKS.md`/`docs/FEATURES.md` one-liners, the `.agents/NOW.md` live-claim row + stamp, and one `.agents/state.md` entry. **NON-COLLISION:** records-only — touches NO model/kernel/registry/loader source, NO CMake, NO test. Co-owns the Kimi-Linear row with `CLAIM-MLA-DEEPSEEK` (MLA half) + `CLAIM-KDA-KERNEL` (KDA host refs), exactly as the K3 row is co-owned by `CLAIM-KIMI-K3-SCOPE`/`CLAIM-KIMI-K3-W2-W5`. | `DONE` | 2026-08-05 — **W0 DEDICATED SPIKE LANDED (records-only, NOT pushed).** Authoritative `config.json` fetched (27 layers = 20 KDA + 7 NoPE-MLA; 256e/top-8/1-shared sigmoid `noaux_tc` `routed_scaling=2.446` `first_k_dense_replace=1`; `num_nextn_predict_layers=0` ⇒ no MTP); reuse-vs-new map with our `file:line`; HW-fit (FITS one GB10, 91.5 GiB / 0.77× pool — real e2e SACRED gate reachable); W0-W7 breakdown + GPU golden-capture recipe. NET-NEW = KDA device kernel (host refs landed, `CLAIM-KDA-KERNEL`), NoPE-MLA branch, hybrid schedule/het-KV, loader name-map. Record checkers green (`check-model-checklist`, `check-agent-record`, `check-doc-checkpoint`, `check-now-current`). |
 | `CLAIM-KIMI-LINEAR-W1` | `MODEL-TEXT-kimi-linear-kimi-linear-for-causal-lm` (stays SPIKE — W1 registry/config/loader scaffolding; forward REFUSES-by-name) | Claude Code (opus-4-8) | isolated worktree `.claude/worktrees/agent-a11ae231a71e8694b`; CPU-only `build-cpu` (`-DVLLM_CPP_CUDA=OFF` Release); NO GPU, NO download — additive TU + one REGISTER line, forward is a `VT_CHECK(false)` stub | branch `worktree-agent-a11ae231a71e8694b`, base `origin/main` HEAD `053116df` | Kimi-Linear-48B-A3B W1 — registry + config + loader + KV-spec scaffolding so the W3-W6 forward can start. Owns ONLY: NEW `include/vllm/model_executor/models/kimi_linear.h`, NEW `src/vllm/model_executor/models/{kimi_linear_registry,kimi_linear_weights,kimi_linear}.cpp`, NEW `tests/vllm/models/test_kimi_linear_scaffold.cpp`, its two CMake registration lines (`CMakeLists.txt` source list + `tests/CMakeLists.txt`), the `KimiLinearForCausalLM` sorted-set + error-message + model-property inserts in `tests/vllm/models/test_model_registry.cpp`, the `MODEL-TEXT-kimi-linear-*` row cells + checklist Status (STAYS `SPIKE`/`📋` — the forward refuses, rollup UNCHANGED), this claim, `.agents/NOW.md`, `docs/STATUS.md`/`docs/BENCHMARKS.md`/`docs/FEATURES.md` one-liners, and one `.agents/state.md` entry. **NON-COLLISION:** additive TU + one REGISTER line ⇒ ZERO edit to any shared array; the forward is a `VT_CHECK(false)` refuse stub so NO production/device path changes; co-owns the Kimi-Linear row with `CLAIM-MLA-DEEPSEEK` + `CLAIM-KDA-KERNEL` + `CLAIM-KIMI-LINEAR-W0`; MUST NOT touch README, Metal/SACRED/apex/darwin, or any other model/kernel source. | `DONE` | 2026-08-05 — **W1 CPU SCAFFOLDING LANDED (foreground, NOT pushed).** Additive registry (`kimi_linear`/`KimiLinearForCausalLM`, `is_hybrid`, `supports_multimodal=false`) + `ParseKimiLinearParams` (20 KDA + 7 NoPE-MLA schedule, MLA 512/128/64/128 q_lora-null, 256e/top-8/1-shared sigmoid `noaux_tc` 2.446, `first_k_dense_replace=1`; asserts `mla_use_nope`/`q_lora==null`) + `EnumerateKimiLinearTensors` name-map VERIFIED vs the real HF safetensors index (MoE = `block_sparse_moe.*`, the correction of the DERIVED K3 `mlp.*`) + loader THROWS BY NAME on missing/mis-shaped tensors + het KV spec (MLA latent-576 + KDA mamba `12288×3`/`32×128×128`) + REFUSE-by-name forward. `test_kimi_linear_scaffold` **9/9·83** + `test_model_registry` **24/24** + `test_kimi_k3_scaffold` 6/6 unchanged; clean CPU Release build; `check-fusion-consistency`/`check-runner-routing-consistency` green (refuse stub skipped); record checkers rc=0. RESIDUAL (W3-W6): KDA device kernel + NoPE-MLA route + sigmoid-noaux MoE + het-KV born-on-runner forward + the e2e SACRED gate (spec §8 recipe). |
 | `CLAIM-KIMI-LINEAR-W2` | `MODEL-TEXT-kimi-linear-kimi-linear-for-causal-lm` (stays SPIKE — W2-W6 CPU REFERENCE forward; the DEVICE runner forward stays refuse-by-name) | Claude Code (opus-4-8) | isolated worktree `.claude/worktrees/agent-aec13af094cd4692f`; CPU-only `build` (`-DVLLM_CPP_CUDA=OFF` RelWithDebInfo); NO GPU, NO download — additive forward TU + a host-materialization pass in the existing loader + additive test | branch `laguna-s21-w7-speed-profile` (worktree), base `origin/main` HEAD `0988db48` | Kimi-Linear-48B-A3B W2-W6 — the CPU REFERENCE forward, per-op gated, replacing the refuse-by-name host `KimiLinearModel::Forward`. Owns ONLY: NEW `src/vllm/model_executor/models/kimi_linear_forward.cpp`, NEW `tests/vllm/models/test_kimi_linear_forward.cpp`, its two CMake lines (`CMakeLists.txt` source list + `tests/CMakeLists.txt`), ADDITIVE host-weight structs + per-op decls in `include/vllm/model_executor/models/kimi_linear.h`, the host-materialization pass in `src/vllm/model_executor/models/kimi_linear_weights.cpp`, the device-refuse doc/rename in `src/vllm/model_executor/models/kimi_linear.cpp`, the `MODEL-TEXT-kimi-linear-*` row cells + checklist Status (STAYS `SPIKE`/`📋` — the DEVICE forward refuses; rollup UNCHANGED), this claim, `.agents/NOW.md`, `docs/STATUS.md`/`docs/BENCHMARKS.md` one-liners, and one `.agents/state.md` entry. **NON-COLLISION:** additive TU + additive header structs; the host `Forward` becomes real but `ForwardDevice` (the DEFAULT gather_logits runner path) STAYS `VT_CHECK(false)` so the born-on-runner CI guard skips it (NO allowlist) and NO production/device path changes; co-owns the Kimi-Linear row with `CLAIM-MLA-DEEPSEEK` + `CLAIM-KDA-KERNEL` + `CLAIM-KIMI-LINEAR-W0/W1`; MUST NOT touch README, Metal/SACRED/apex/darwin, or any other model/kernel source. | `DONE` | 2026-08-05 — **W2-W6 CPU REFERENCE FORWARD LANDED (foreground, NOT pushed).** Real host `KimiLinearModel::Forward` composes the whole 27-layer hybrid from the landed primitives: KDA layer (`vllm::kimi_kda` refs + the gated-delta recurrence `fused_recurrent.py:122-149`), NoPE-MLA (materialized-MHA reference, scaling `qk_head**-0.5`, no RoPE), sigmoid `noaux_tc` MoE (bias-select/unbiased-weight, renormalize, `routed_scaling=2.446`) + shared expert, dense layer-0 SwiGLU; loader now MATERIALIZES the host float weights (bf16/f32→f32). `test_kimi_linear_forward` **6/6·246** (KDA==kimi_kda composition; NoPE-MLA==materialized-MHA ref; router+block==hand-computed; loader materializes; whole 2-layer forward finite/coherent; greedy-decodes 5 tokens, context advances) + `test_kimi_linear_scaffold` 9/9·83 + `test_kimi_kda` 14/14 UNCHANGED; clean CPU build; `check-fusion-consistency`/`check-runner-routing-consistency` green (DEVICE forward refuse-skipped, no allowlist). RESIDUAL: the DEVICE born-on-runner forward + the W0/W7 e2e SACRED token golden on GB10 (spec §8 recipe) + speed. |