Will Hampson d03abf43d3 Fix estate GPU visibility overrides (#447) ...

Generate a per-instance Compose override for estate boots so CUDA_VISIBLE_DEVICES and NVIDIA_VISIBLE_DEVICES are set inside each container. This prevents parallel vLLM estate instances from both binding to physical GPUs 0/1 when Docker exposes all reserved devices.\n\nValidated with test-diagnose-estate, test-estate-json, test-profiles-compat, test-launch-compat, and a live 4x3090 estate boot.

2026-06-21 04:38:46 +05:00

..

test-bench-agentic-ramp.sh

feat(bench-agentic): decouple context ramp from tool-call success (#255) (#396)

2026-06-13 18:53:54 +05:00

test-classifier.sh

feat(pull): v0.8.2 STEP V1 — capture-on-hard-block pt1-gate emitter + BaseCaptureBundle protocol lift

2026-05-18 16:43:28 +00:00

test-compose-image-drift.sh

refactor(vllm): reconcile vLLM engines to v0.22.0 — two-engine split (#254)

2026-06-05 11:27:17 +00:00

test-compose-mounts-resolve.sh

Prune Gemma vLLM variants (9->3) and pin survivors to v0.21.0

2026-05-29 07:59:42 +05:00

test-compose-nvlink-escape.sh

Add guard test: conditional compose entrypoints must $$-escape runtime bash

2026-06-18 17:45:45 +00:00

test-compose-registry-disk.sh

Add llamacpp/qwen27b-pi-reasoning-single (Qwen3.6-27B Pi-style coding agent, mainline llama.cpp + MTP) (#425)

2026-06-18 03:56:28 +05:00

test-compose-restart-policy.sh

Operational robustness (#281): orphan-safe switch.sh · reboot-surviving vLLM · multi-GPU power sweep (#285)

2026-05-31 22:13:52 +05:00

test-compose-status-drift.sh

feat(registry): slug health/availability flag (#265)

2026-05-30 21:10:49 +05:00

test-dedup.sh

feat(pull): v0.8.2 STEP V1 — capture-on-hard-block pt1-gate emitter + BaseCaptureBundle protocol lift

2026-05-18 16:43:28 +00:00

test-default-resolver.sh

refactor(gemma): rename dual slugs + ladder gemma-bf16-mtp to 131K (#286)

2026-06-01 01:20:46 +05:00

test-diagnose-estate.sh

Fix estate GPU visibility overrides (#447)

2026-06-21 04:38:46 +05:00

test-diagnose-profile.sh

feat(gemma): gemma duals → vLLM v0.22.0 (rebase #40391 lean + re-instate #42006) (#287)

2026-06-01 01:21:51 +05:00

test-estate-json.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-generate-compose.sh

Fix Genesis cleanup test fixtures

2026-06-05 14:02:38 +00:00

test-generate-from-profile.sh

Repoint vllm/gemma-26ba4b-single to INT8-PTH long-ctx (#465)

2026-06-06 11:30:27 +00:00

test-gpu-mode-list.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-health-container.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-hf-repos-resolve.sh

test(weights): opt-in HF-repo resolve guard (#320) (#409)

2026-06-15 06:35:39 +05:00

test-hwdetect.sh

feat(pull): v0.8.2 STEP V4 — optional whichllm hw-detect subprocess (CONTRACT-3, hw-detect-only)

2026-05-18 19:29:43 +00:00

test-kv-calc-fit.sh

Add kv-calc --fit-all: batch fit for all registry slugs in one process

2026-06-18 18:34:21 +00:00

test-kv-generic-dense.sh

Wire gemma-4-26b-a4b AWQ on stock vLLM v0.22.0; retire AutoRound (#326)

2026-06-06 00:46:13 +00:00

test-kvcalc-version.sh

v0.8.0 [F] F6: CONTRACT-5 mandatory content-hash kv_calc_version (G2) + G1 topo-verify + L2 fixture sync

2026-05-17 04:50:50 +00:00

test-launch-compat.sh

chore(327): archive the Qwen3.6-27B DFlash path + deprecate vllm-nightly-dflash

2026-06-05 15:26:29 +00:00

test-launch-registry-parity.sh

feat(registry): slug health/availability flag (#265)

2026-05-30 21:10:49 +05:00

test-list-topology-filter.sh

chore(327): archive the Qwen3.6-27B DFlash path + deprecate vllm-nightly-dflash

2026-06-05 15:26:29 +00:00

test-loop-input.sh

Prune Gemma vLLM variants (9->3) and pin survivors to v0.21.0

2026-05-29 07:59:42 +05:00

test-measurement-record.sh

test-measurement-record: make BENCH_MOCK capture hermetic (#478)

2026-06-10 00:47:10 +00:00

test-model-default-resolver.sh

chore(327): archive the Qwen3.6-27B DFlash path + deprecate vllm-nightly-dflash

2026-06-05 15:26:29 +00:00

test-model-weights-registry.sh

fix(35b-a3b): correct AutoRound INT4 hf_repo (#316)

2026-06-05 06:36:15 +00:00

test-parallel-boot.sh

feat(estate): add parallel boot mode

2026-05-15 22:18:09 +05:00

test-patch-attribution.sh

feat(pull): v0.8.2 STEP V3 — arch-registry expansion + chat-template attribution/drift_guard

2026-05-18 18:29:58 +00:00

test-power-cap-sweep-multigpu.sh

Operational robustness (#281): orphan-safe switch.sh · reboot-surviving vLLM · multi-GPU power sweep (#285)

2026-05-31 22:13:52 +05:00

test-preflight-compose-deps.sh

ik-llama: migrate spec-dec flags to --spec-type + digest-pin the image (#401)

2026-06-14 00:21:35 +05:00

test-preflight-vram.sh

Add hardware-aware compose preflight

2026-05-13 18:03:37 +00:00

test-profiles-compat.sh

Add opt-in LMCache KV-offload compose (vllm/qwen-27b-dual-lmcache, incubating) (#421)

2026-06-17 20:46:38 +05:00

test-pull-swap.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-pull.sh

chore(327): archive the Qwen3.6-27B DFlash path + deprecate vllm-nightly-dflash

2026-06-05 15:26:29 +00:00

test-pullemit-capture.sh

Wire gemma-4-26b-a4b AWQ on stock vLLM v0.22.0; retire AutoRound (#326)

2026-06-06 00:46:13 +00:00

test-pullgate-deriver.sh

refactor(compose): insert <quant> layer + make the registry the single source of truth

2026-05-26 15:58:08 +00:00

test-pullgate-download.sh

v0.8.0 [E] E2-fix-2: verify *.safetensors against HF API lfs.sha256 (not Xet-redirect-fragile HEAD x-linked-etag) — on-rig E5 caught false no-etag

2026-05-17 01:28:22 +00:00

test-pullgate-gates.sh

Pull gate: actionable messages for uncurated derives of curated models (#424)

2026-06-18 01:10:01 +05:00

test-quality-baseline.sh

Add curated quality-baseline corpus + auto-diff (#252 Phase 1) (#397)

2026-06-13 19:11:35 +05:00

test-quality-thinking.sh

quality-test / rebench-full: forward --max-tokens to benchlocal-cli (#426)

2026-06-18 05:57:19 +05:00

test-registry-json.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-report-calib.sh

chore(35b-a3b): drop the built-in-MTP A/B compose (vllm/qwen-a3b-preview-mtp) (#262)

2026-05-30 17:52:09 +05:00

test-setup-picker.sh

Fix Genesis cleanup test fixtures

2026-06-05 14:02:38 +00:00

test-stream-toolcall-probe.sh

verify-full: add streaming tool-call check + a reusable streaming probe (#404)

2026-06-14 04:17:04 +05:00

test-submit-bench.sh

feat(eval): expose request-level thinking toggles (#196)

2026-05-23 08:24:58 +05:00

test-submit-pull.sh

Wire gemma-4-26b-a4b AWQ on stock vLLM v0.22.0; retire AutoRound (#326)

2026-06-06 00:46:13 +00:00

test-switch-explain.sh

Phase 2b: add --json/CLI contracts to 7 stack scripts (additive)

2026-06-18 08:43:00 +00:00

test-switch-orphan-teardown.sh

Operational robustness (#281): orphan-safe switch.sh · reboot-surviving vLLM · multi-GPU power sweep (#285)

2026-05-31 22:13:52 +05:00

test-switch-registry-parity.sh

gemma-4-12b: register single-card slugs beellama/gemma-12b + llamacpp/gemma-12b

2026-06-04 05:37:16 +00:00

test-trust-pipeline.sh

refactor(compose): insert <quant> layer + make the registry the single source of truth

2026-05-26 15:58:08 +00:00

test-verify-stress-ceiling.sh

fix(verify-stress): three live-caught bugs in ceiling ladder (#199)

2026-05-23 04:42:18 +00:00

test-weights-revision.sh

feat(weights): optional revision: pin in weights-fetch schema (#319) (#408)

2026-06-15 06:35:43 +05:00