# FIX Antenna .NET: both receiver-placement DAC campaigns complete

Both accepted rigs complete four functional and twelve full report pairs per
host, with all 32 original role exits zero. Both independent cold/admission
workers replay all 120,080 original primary and scheduler-delay values per host
and refuse 107 actual-output or metadata corruptions. Both bounded collectors
also pass: 969,582 bytes from HP and 970,569 bytes from SCAN, with zero full
sample-array or clock-table downloads. Every original SDK/session, canonical
prefix/administration, complete typed owner, source/clock/sample, namespace/UID,
CPU, observed host/NIC and physical-wire gate passes.

The separately qualified original receiver thread stays on its selected CPU
inside the inherited role mask, from explicit unmeasured preflight through
original Logout. Owner/receiver CPUs remain HP 8/9 and 13/14, SCAN 10/11 and
14/15. Both thirteen-file immutable deployments, actual loaded original SDK/
helper/harness, parsed invocation, native receiver TIDs and complete pin/closure/
phase timestamp proof remain bound. No SDK, producer or harness is rebuilt for
timing and no processor-count override, custom transport, root helper/sudo,
shared-harness or old output is changed.

The report contract remains 1,000 warmup, 10,000 measured operations, three
repeats and 10,000/s. All prefix, warmup/prewarm and typed owner/sample partitions
pass, with zero measured application-frame copies. Every flow still fails rate
and scheduler delay. The independently reproduced stability decisions and full
actual diagnostic rate ranges follow; none is an admitted comparison number.

| Rig | Flow | Actual report rate range (/s), diagnostic | Stability failure |
|---|---|---:|---|
| HP | `nos_er` | 3,837.859–3,924.561 | no |
| HP | `md` | 3,620.538–3,701.395 | no |
| HP | `md_nos_er` | 2,772.542–2,829.324 | no |
| HP | `tick_to_trade` | 5,903.902–6,097.721 | no |
| SCAN | `nos_er` | 3,747.482–3,818.462 | no |
| SCAN | `md` | 3,561.203–3,602.784 | yes |
| SCAN | `md_nos_er` | 2,814.349–2,992.610 | yes |
| SCAN | `tick_to_trade` | 3,455.809–3,838.593 | yes |

Tick also retains the explicit SDK-envelope prepared-application boundary
blocker. Receiver placement does not turn the original SendMessage API into a
completed-wire boundary and does not substitute its after-write callback for t0.
Both compact admission writers execute and write **zero cells**. All retained
good result keys stay unchanged; no matched .NET admission or lower-rate waiver
is inferred. The source-specific campaign is complete and its unchanged failed
matrix is not rerun. This establishes no matched performance improvement over
the earlier separately recorded receiver configuration.

The physical request holds the outer quiet `isol-no-shared` lease, at high
priority on HP and normal priority on SCAN. Before/after observed host facts
match and all physical transmit deltas are positive. Preparation and independent
functional/cold work use their isolated leases. Fifty physical/writer source
controls, 144 immutable-runtime controls and 73 original placement/preflight
controls pass. An added strict receiver-CPU writer model initially lacks its
synthetic endpoint map; the corrected model supplies it and all fifty focused
checks pass, without any native pair or physical rerun.

| Binding | Source / SHA-256 |
|---|---|
| Original receiver producer source | `5c5aa6675e76df00ce075c83cda1b12dad169ccb` |
| Immutable deployment source | `2fe19e1c7acd6e6c71b379ab166118cbb4ab2915` |
| Physical/admission controller source | `a39042c13dc5ab98c991f6e7cbfa3ba7fc52864a` |
| Original receiver DLL | `f2cd673a4967d8e2c7acaac95862475f3603584808372643b791896af8f5e481` |
| HP completed.json | `176dfab49e60815ef39f8e62de33f3deb5f0f06450a5d5f7247571ea1a8598da` |
| HP independent-replay.json | `b4c305a9f5b3911f21c1ecb98bf45a3e7247c4d70c8f1362abae014a7235e7d8` |
| HP result-admission.json | `b766d9814c0e4f354d2cc96b1112146d5ec6dc2e8bad739d7921efaf9a72a4a0` |
| HP reviewed.json | `da8d7f1304f4752292a5e95f13b86843f5d06b8fab6e64993932a6476344962f` |
| SCAN completed.json | `83edbbdb66649f2e82daf40020a16385d731086de6a864413af6e1efe80f0fd8` |
| SCAN independent-replay.json | `3f36cdf0ddd12ecfb41e79dd8f89fd9ad4e73a80e43b14672d89528cdf9f5111` |
| SCAN result-admission.json | `6dfab235cc47451a83d3eac28a787c028be0de540fdfd626739004091660c15e` |
| SCAN reviewed.json | `9592489333cd74e52b95c1100b976b886192adeca1d2b2a40ec83b3276ad9b25` |

All original working output remains under
`~/libhft-bench-runs/fixantenna-receiver-physical-20261026-v1` on its host;
it is not manually pruned. Complete independent bounded proofs are under
`~/build/codex-fullbench/libhft/fixantenna-receiver-physical-review-20261026-v1/{hp,scan}`.
Both `codex-fullbench-fixantenna-receiver-physical-20261026-v1.service` and
both `codex-fullbench-fixantenna-receiver-admission-20261026-v1.service`
units exit successfully. See the [full receiver qualification](RECORD_2026-10-26_fixantenna_receiver_placement.md),
[earlier original report campaign](RECORD_2026-10-26_fixantenna_physical.md)
and [separate CPU investigation](RECORD_2026-10-26_fixantenna_measured_cpu.md).