Default to Gyro-S, record B580 test resultsHEADmaster

Gyro-M does not fit a 12 GB B580 with 31 GiB RAM on the xe
stream-experts path: the engine refused with 5,309 experts neither
in VRAM nor in the pinned RAM mirror. Gyro-S fits and answers at
22-29 tok/s, but hit UR_RESULT_ERROR_DEVICE_LOST in the verify step
on 2 of about 6 requests.

AGENTS.md records the upstream fake-RAM bug that adds --ple-io ram
to Gyro configs, the measurements, the suspected spin-bound cause
and the untried fixes. Logs are now ignored.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

1 files changed, 32 insertions, 0 deletions

diff --git a/AGENTS.md b/AGENTS.md
index caddcd5..5d97a07 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -31,9 +31,41 @@ reimplement anything upstream's scripts already do.
   12 GB card, so close other programs before starting it.
 - Gyro on Arc was tested upstream only on an A750 (`i915`). The B580 is
   untested natively; one WSL2 B580 report saw device loss.
+- Measured here (v0.2.2, B580, 31 GiB RAM): on `xe` the engine runs with
+  `--stream-experts` and `STRATA_VERIFY_NO_HOST`, so every expert must sit in
+  the VRAM cache or a pinned RAM mirror (capped at MemAvailable minus 4 GiB).
+  Gyro-M does not fit: 4.6 GiB VRAM cache + 21.6 GiB mirror left 5,309
+  experts out and the engine REFUSED. setup's "35 GB of VRAM + RAM: fits"
+  check is wrong for this path. Gyro-S fits: 3,062 experts in VRAM, 21,514
+  mirrored (21 GiB), about 6 GiB of RAM left, decode 22-29 tok/s.
+- Gyro-S is not stable here: 2 of about 6 requests hit
+  `UR_RESULT_ERROR_DEVICE_LOST` in the verify (speculative decoding) step,
+  the engine segfaulted (139) and the chat showed "verify error". The server
+  restarts the engine and the card recovered each time.
+- Likely cause, not verified: device spins on host flags are capped at
+  20,000 reads on B-series cards (`sycl/include/strata/sycl_doorbell.hpp`),
+  tuned on a B70 that holds every expert in VRAM. Here 21 GiB of experts
+  come from the RAM mirror over PCIe, so a wait can outrun the cap. Upstream
+  documents the same 139 crash pattern for a cold first request
+  (`docs/INTEL.md`, `STRATA_SYCL_SPIN_MAX`).
+- Untried next steps: (1) `"env": {"STRATA_SPIN_MAX": "2000000"}` in the
+  config, read at engine start. Risk: on `xe` an overlong spin can wedge the
+  GPU until reboot, and the B580 drives the desktop. (2) Drop `--spec`,
+  `--spec-min-p` and `--mtp` from the config to avoid the draft verify, at
+  lower decode speed.
+- Status 2026-10-22: testing stopped. Gyro-M deleted. Gyro-S (55 GiB) and
+  the shared draft layer remain in `Strata-data/`.
 - On `xe`, asking for more VRAM than the card has does not fail: buffers
   spill into RAM and the host can livelock. Never raise `--context` or lower
   `--vram-reserve-mib` past what setup's fit rule picked.
+- Upstream bug (v0.2.2): `setup_intel.py` tells `setup.py` the host has
+  1024 GB of RAM (`S.ram_gb = lambda: fake_ram`), so on the `xe` path
+  `setup.py` adds `--ple-io ram` to the Gyro config. That loads the 54 GB
+  n-gram table into RAM and would OOM this host. The Intel code itself says
+  Gyro's table must stay on the SSD (it only guards the `i915` branch). After
+  every `setup` run, check `Strata/strata-gyro-gyro-m.json` and drop
+  `--ple-io ram` from `args`. Start the model with
+  `Strata/run-gyro-gyro-m.sh`, which does not rewrite the config.
 - Pictures are off by default on Battlemage (they reset a B70). Do not pass
   `--vision yes`.
 - `docker build --pull` matters: a stale base can carry an ocloc that crashes