# round 22 track b: router + glue

## delivered

patch: `/tmp/sol_r22_router_glue/sol_r22_router_glue.patch`
sha256: `87651f461c679bb12d1decbe0f954172d04cda6ca3730f50fcbae6cc59fd3ef4`

only `metal/kernels/glue.metal` changes. tracked vendor sources were not edited. `git -C vendor/strata apply --check` exits 0. baseline vendor head: `296b59d2f73c8f04a805df29c2fa97655adace78`; baseline actual kernel snapshots are in `/tmp/sol_r22_router_glue/baseline/` (the checkout may contain prior changes, so snapshots, not head alone, define this experiment).

`swiglu_prep` moves scale tables to program-scope constants and replaces its 16-float threadgroup array/barrier with same-order shuffle sums. lane 0 still adds even indices and odd indices separately, then adds the two results. all threads execute the shuffle, outside the lane-0 output branch. production dispatch uses 16 threads, entirely within one simdgroup. pipeline-reported static threadgroup memory changes **64 -> 0 bytes**; source has **one -> zero threadgroup barriers**. maximum threads remains 1024. no interface/layout or host dispatch changes.

**no measured latency win established.** the parent explicitly requested retaining this resource-only reduction after seeing initial results. it is not an end-to-end throughput claim.

## rejected hypotheses

- full `route_body` unrolling plus constant-index winner invalidation: byte-identical, but consistently slower. confirmed paired median regressions are below. reverted completely, including the entry barrier guard.
- router entry barrier guard alone: +0.3% to +3.8% paired median latency across both formats/windows, not a win. not retained.
- combine `n_embd == 2560` constant-division fast path: approximately -1.2% to -0.3%, below the predeclared 3% significance threshold. not retained.
- combine simd-broadcast plan entries and shared sigmoid: +6.2%, +2.2%, +2.6%, and +0.7% for ntok 1/2/4/8, respectively. not retained. this candidate was only evaluated for width 2560, where groups do not cross token boundaries; it was not generalized to arbitrary widths.

no disassembly/register-count evidence was collected. increased register pressure / occupancy loss is a plausible explanation for the unroll regression, **not a proven compiler diagnosis**. source unrolling does not establish zero stack traffic or prove the original array spilled.

## correctness

- `fuse_parity`: exit 0, 5.664s, **all pass**. includes bf16/f32 router logits/ids/w/plan bitwise checks for ntok 1..8 and repeated counter reset; fused norm/combine regressions also pass.
- `route_parity`: exit 0, 2.180s; 200 k=512 and 200 k=256/moce windows, zero mismatches. this existing cpu-reference test uses a floating-weight tolerance, not memcmp; it is supplementary, not the byte-parity evidence.
- custom paired harness: full output-buffer `memcmp`, sentinel-filled padding, 8 trials per kernel/window; ntok 1/2/4/8. **every comparison passed** for logits, ids, w, plan, xp and bo. counter reset also checked. trial zero deliberately produces exact tied router logits; later trials vary activations. router trials alternate 256/512 experts, with/without moce and draft scaling, random-like fixed residency; rejected full unroll additionally passes rank-1 scale 2 specialization.
- delivered source rechecked for router_route_30/router_route_0/moe_combine_sh (`delivered_*.log`, unchanged kernels) and swiglu (`production*.log`, changed kernel). swiglu tests the actual width **640**, `a.n_tok=8` plane stride, and dispatched windows 1/2/4/8; unused rows remain sentinel-filled and match too. early exploratory swiglu runs used width 2560 and are not the production-shape evidence.

stock fuse compilation exceeded two original 4.9s deadlines. after explicit parent authorization for a 20s fuse timeout, its next run passed in 5.664s. all custom benchmark invocations stayed below 5s. no full model, no port 8080 interaction, no subagents.

## paired gpu timings

apple m5; runtime safe math; release appleclang 21.0.0.21000327. same process/library/inputs, separate outputs. final protocol: 0.10s alternating-kernel warmup, ten pairs of 256 dispatches, alternating which side runs first. table reports median baseline/candidate microseconds and median paired percentage (not ratio of medians). clocks/shared-gpu noise are visible in exploratory logs; unchanged combine control differs by about 1%, so changes near that scale are not credible wins.

production swiglu, three independent short invocations/window, 30 pairs pooled:

| kernel | ntok | base us | candidate us | paired change |
|---|---:|---:|---:|---:|
| `swiglu_prep` | 1 | 1.898 | 1.882 | -0.94% |
| `swiglu_prep` | 2 | 1.957 | 2.020 | +0.79% |
| `swiglu_prep` | 4 | 2.519 | 2.583 | +1.14% |
| `swiglu_prep` | 8 | 3.742 | 3.776 | +1.19% |

rejected full-unroll router, default k=512/no moce timed path (the parity trials still cover moce):

| kernel | ntok | base us | candidate us | paired change |
|---|---:|---:|---:|---:|
| `router_route_30` | 1 | 42.309 | 43.717 | +9.86% |
| `router_route_30` | 2 | 42.039 | 50.012 | +16.05% |
| `router_route_30` | 4 | 49.476 | 54.251 | +10.32% |
| `router_route_30` | 8 | 93.427 | 103.733 | +9.35% |
| `router_route_0` | 1 | 65.351 | 71.745 | +10.63% |
| `router_route_0` | 2 | 77.906 | 90.182 | +13.80% |
| `router_route_0` | 4 | 81.686 | 87.506 | +6.23% |
| `router_route_0` | 8 | 89.646 | 82.581 | +7.30% |

unchanged combine control, delivered source:

| kernel | ntok | base us | candidate us | paired change |
|---|---:|---:|---:|---:|
| `moe_combine_sh` | 1 | 9.784 | 9.888 | -0.05% |
| `moe_combine_sh` | 2 | 4.078 | 3.927 | -6.02% |
| `moe_combine_sh` | 4 | 1.892 | 1.893 | -1.64% |
| `moe_combine_sh` | 8 | 7.574 | 7.638 | +0.37% |

## reproduction / artifacts

all experimental sources, baseline snapshots, failed attempts and logs: `/tmp/sol_r22_router_glue/`; source hashes: `source_sha256.json`. `plan.md` records the predeclared threshold. `ab.metal` is the delivered baseline+candidate library; `ab_full_unroll.metal`, `ab_barrier_only.metal`, `ab_combine_broadcast.metal` retain rejected libraries. `ab_initial.cpp`, `ab_v2.cpp`, `ab_full_confirmation.cpp` preserve earlier evaluators; `ab.cpp` is the delivered evaluator.

```text
cmake -S metal -B build -G Ninja -DCMAKE_BUILD_TYPE=Release
cmake --build build --target fuse_parity route_parity -j 4
# commands above run in /tmp/sol_r22_router_glue/strata
c++ -std=c++20 -O2 -I/tmp/sol_r22_router_glue/strata/metal/src /tmp/sol_r22_router_glue/ab.cpp /tmp/sol_r22_router_glue/strata/build/libmtl_rt.a -framework Metal -framework Foundation -o /tmp/sol_r22_router_glue/ab
/tmp/sol_r22_router_glue/ab swiglu_prep 1
# repeat ntok 2/4/8; other kernels: router_route_30, router_route_0, moe_combine_sh
/tmp/sol_r22_router_glue/strata/build/fuse_parity
/tmp/sol_r22_router_glue/strata/build/route_parity
git -C vendor/strata apply --check /tmp/sol_r22_router_glue/sol_r22_router_glue.patch
```

current `Window::comb_push` is at 1083–1091, not the brief's stale 1015–1045 range; no host changes were needed. final image review completed for the entire diff and surrounding glue code; scale-table placement and shuffle idiom match neighboring `gr.metal`.