feat(agentx) Use Magpie native AgentX for Hyperloom optimization by haofrank · Pull Request #1675 · AMD-AGI/Hyperloom

haofrank · 2026-09-25T05:41:27Z

Fresh HYPERLOOM_AGENTX=1 sessions use Magpie-managed AgentX through the existing Hyperloom CLI. Hyperloom keeps optimization, candidate grading and selection. Magpie manages the server lifecycle and optional trace capture; InferenceX/AIPerf replays and measures the workload. No Slurm integration is required.

Fixes #1601.

What this enables

Single-node SGLang/vLLM sessions automatically use the native backend. Hyperloom translates model, image, topology and concurrency settings into Magpie configuration.
Server arguments, environment tuning and verified source/kernel changes reach the actual process. Hyperloom checks the launch receipt while keeping workload identity fixed.
KEEP/resume and GEAK remeasurement remain in Hyperloom. GEAK proposals require review and canonical remeasurement; proxy scores cannot promote a candidate.
Historical sessions retain their recorded backend, measurement contract and accepted records. New measurements still face current validity checks. Synthetic and explicit MLPerf paths remain available.

Compatibility and trace handling

Successful standalone diagnostic profiles now reach the Coordinator's trace-analysis path. Failed or incomplete captures are rejected, and diagnostic results cannot change baseline/KEEP evidence.
The installer and CLI share one Magpie health check: generic sessions require an importable package; native AgentX additionally requires the pinned capabilities, commit and audited execution tree. Compatible custom/offline installs and saved legacy sessions retain their existing behavior.
Native CLI preflight rejects non-empty legacy AIPERF_*, AGENTIC_*, AGENTX_DATASET, AGENTX_WARMUP_REQUESTS_PER_LANE and WEKA_LOADER_OVERRIDE controls in the shell or benchmark.envs, before dependency installation or benchmarking. Errors identify the variable names without printing their values. Use supported benchmark.agentx settings and the resolved InferenceX recipe. Legacy and MLPerf controls remain available.

Profiling

Hyperloom profiles an accepted configuration using Magpie's AgentX replay. The corresponding Magpie settings can request repeated captures:

benchmark:
  envs:
    CONC: 32
  profiler:
    torch_profiler:
      enabled: true
      num_steps: 20
      num_profiles: 3
      start_seconds: 30
      interval_seconds: 200
      detailed_annotations: false

Use the accepted workload's concurrency. The initial delay starts after warmup, and later intervals start after the previous trace finishes flushing. These are server execution steps, not request counts. Profile results are diagnostic; KEEP requires separate unprofiled canonical evidence. Rank-file inventory and selected-rank quality certificates remain distinct.

The example uses basic tracing. Enhanced shape/roofline analysis requires a compatible instrumented runtime and passing trace-quality checks.

Dependencies and validation

Hyperloom d9bb74c pins merged Magpie d80eb4d from #117, with its updated audited package fingerprint and InferenceX 408c015. SGLang shape discovery is optional; per-batch graph trace export is preserved. Both InferenceX layouts are supported. The generic packaged throughput client has an isolated Python 3.12 / Transformers 5 runtime; native AgentX/AIPerf remains unchanged. The validation harness protects GSM8K's existing environment with framework constraints and version checks. Generic compatibility preparation leaves the audited Magpie package unchanged.

At dd02d4ea3, 1,609 focused regression tests passed (1 skipped), including Coordinator trace promotion, installer/preflight compatibility, candidate execution, profiling, Roofline and GEAK boundaries. Ruff lint/format and shell syntax checks passed. The shared native health check also passed against the actual VCS-installed Magpie d80eb4d package. Main dc46f8d97 is merged; conflict resolutions follow its relocated trace/GEAK helpers and retired recipe-rewrite path. Current-head CI is pending; local mypy was unavailable.

Before these fixes, d4ccb156a passed full Linux Python 3.10/3.11 unit and coverage CI. Its separate single-GPU Qwen3-0.6B E2E run timed out and did not establish a passing E2E result.

GPU baseline/candidate and trace capture previously passed on earlier revisions; the slower --disable-cuda-graph candidate was not promoted. Selected-rank trace quality reported blocked/warn, while other ranks received file-inventory checks. Full TraceLens/Roofline analysis has not run. Current-head native GPU, ordinary benchmark/GSM8K, and the full Coordinator/KEEP/GEAK GPU loop remain unverified. See detailed results and limitations.

Copilot

Copilot review overview

🟡 Changes recommended

Compatibility checkout preparation blocks the async event loop, and the native GEAK skip leaves superseded handoff code reachable only through direct tests.

Review effort: Balanced
Findings: 2

Open (2)

Run synchronous AgentX checkout in a worker thread · New
Remove unreachable native AgentX handoff and proxy path · New

What changed in this PR

Routes AgentX measurements through pinned Magpie/InferenceX native launchers while preserving recipe topology, cache behavior, and measurement provenance.

Changes:

Adds native AgentX configuration, runtime validation, topology-aware metrics, and resume state.
Rejects unsupported optimization, GEAK, mutation, and sweep paths.
Updates installation, profiling compatibility, tests, examples, and documentation.

File	Description
`src/hyperloom/orchestrator/state/shared_state.py`	Persists AgentX pins and source config.
`src/hyperloom/orchestrator/phases/prelude.py`	Preserves pinned InferenceX checkout.
`src/hyperloom/orchestrator/phases/kernel.py`	Skips unsupported native GEAK execution.
`src/hyperloom/orchestrator/loop/writeback.py`	Persists AgentX GPU topology.
`src/hyperloom/orchestrator/kernel/request_handlers.py`	Rejects native mutation integration.
`src/hyperloom/orchestrator/kernel/conc_sweep.py`	Records AgentX GPU count.
`src/hyperloom/orchestrator/kernel/conc_sweep_plot.py`	Normalizes plots by recipe GPUs.
`src/hyperloom/orchestrator/actions/executors/tests/test_integrate_patch_executor.py`	Tests native integration rejection.
`src/hyperloom/orchestrator/actions/executors/profile.py`	Adds isolated compatibility profiling.
`src/hyperloom/orchestrator/actions/executors/integrate_patch.py`	Blocks native source mutations.
`src/hyperloom/orchestrator/actions/executors/baseline.py`	Adds native launch validation hooks.
`src/hyperloom/orchestrator/actions/executors/_recipe_script.py`	Resolves native server proxies.
`src/hyperloom/orchestrator/actions/executors/_ray_backend.py`	Forces AgentX local execution.
`src/hyperloom/orchestrator/actions/executors/_inferencex_patcher.py`	Limits patching to explicit checkouts.
`src/hyperloom/orchestrator/actions/executors/_grid_runner.py`	Validates native grid topology.
`src/hyperloom/orchestrator/actions/executors/_grid_base.py`	Adds GPU-count result metadata.
`src/hyperloom/orchestrator/actions/executors/_benchmark_runtime.py`	Preserves native launcher selection.
`src/hyperloom/orchestrator/actions/executors/_accuracy_gate.py`	Parses native request-error gates.
`src/hyperloom/inference_optimizer/tests/test_workload_envs.py`	Tests diagnostic profile materialization.
`src/hyperloom/inference_optimizer/tests/test_workload_envs_branches_unit.py`	Tests native configuration safeguards.
`src/hyperloom/inference_optimizer/tests/test_warm_replay.py`	Tests pinned checkout preservation.
`src/hyperloom/inference_optimizer/tests/test_shared_state_units.py`	Tests disabled AgentX warm replay.
`src/hyperloom/inference_optimizer/tests/test_run_optimize_exit_code_lock.py`	Strengthens CLI error assertions.
`src/hyperloom/inference_optimizer/tests/test_ray_backend_unit.py`	Tests local execution and GPU counts.
`src/hyperloom/inference_optimizer/tests/test_profile_and_kernel_handlers.py`	Tests isolated compatibility profiling.
`src/hyperloom/inference_optimizer/tests/test_map_aiperf.py`	Tests native accounting mapping.
`src/hyperloom/inference_optimizer/tests/test_kernel_integrate_and_report.py`	Tests measurement-only integration behavior.
`src/hyperloom/inference_optimizer/tests/test_install_magpie_idempotent.py`	Tests Magpie capability repair.
`src/hyperloom/inference_optimizer/tests/test_install_inferencex.py`	Tests checkout ownership preservation.
`src/hyperloom/inference_optimizer/tests/test_inferencex_preflight_clone.py`	Tests pinned checkout identity.
`src/hyperloom/inference_optimizer/tests/test_inferencex_patcher.py`	Tests isolated patch scope.
`src/hyperloom/inference_optimizer/tests/test_grid_runner.py`	Updates legacy timeout coverage.
`src/hyperloom/inference_optimizer/tests/test_graded_comparison.py`	Tests native grading guards.
`src/hyperloom/inference_optimizer/tests/test_geak_revalidation_dispatch.py`	Tests GEAK refusal and recovery.
`src/hyperloom/inference_optimizer/tests/test_geak_handoff_workload_contract.py`	Tests pre-handoff GEAK skipping.
`src/hyperloom/inference_optimizer/tests/test_geak_handoff_launch_server_script.py`	Tests native proxy resolution.
`src/hyperloom/inference_optimizer/tests/test_eval_result_dir_wiring.py`	Isolates native accuracy mapping.
`src/hyperloom/inference_optimizer/tests/test_env_safety.py`	Tests blocked AgentX identity overrides.
`src/hyperloom/inference_optimizer/tests/test_coordinator_gemm_promote_units.py`	Separates generic grading tests.
`src/hyperloom/inference_optimizer/tests/test_cli_conc_sweep_default.py`	Tests mode-dependent sweep defaults.
`src/hyperloom/inference_optimizer/tests/test_cli_bootstrap.py`	Tests source and pin snapshots.
`src/hyperloom/inference_optimizer/tests/test_baseline_param_overrides.py`	Tests unsupported native overrides.
`src/hyperloom/inference_optimizer/tests/test_backend_gating.py`	Tests immutable install behavior.
`src/hyperloom/inference_optimizer/tests/test_agentx_runtime.py`	Tests native launch-boundary validation.
`src/hyperloom/inference_optimizer/tests/test_agentx_repair.py`	Updates installer extraction coverage.
`src/hyperloom/inference_optimizer/tests/test_agentx_baseline_timeout.py`	Tests profile compatibility timeout.
`src/hyperloom/inference_optimizer/tests/test_accuracy_gate_units.py`	Tests native error thresholds.
`src/hyperloom/inference_optimizer/cli/parser.py`	Adds benchmark-config CLI semantics.
`src/hyperloom/inference_optimizer/cli/bootstrap.py`	Seeds native state and runtime pins.
`src/hyperloom/inference_optimizer/baseline_comparison/research_hints.py`	Uses physical recipe GPU count.
`src/hyperloom/inference_optimizer/assets/install.sh`	Pins and validates Magpie 0.3.0.
`src/hyperloom/inference_optimizer/assets/agentx/aiperf_client.sh`	Uses canonical AgentX model identity.
`src/hyperloom/inference_optimizer/agentx/mapping.py`	Maps native accounting fields.
`src/hyperloom/common/tests/test_perf_metric.py`	Tests per-chip throughput guards.
`src/hyperloom/common/perf_metric.py`	Adds topology-aware throughput helpers.
`src/hyperloom/common/gain_math.py`	Normalizes sweep guards per chip.
`src/hyperloom/common/env_safety.py`	Blocks recipe-identity candidate overrides.
`examples/hyperloom-agentx-glm-5.2-mi355x/README.md`	Documents the runnable GLM example.
`examples/hyperloom-agentx-glm-5.2-mi355x/benchmark.yaml`	Adds the native GLM recipe config.
`docs/release-notes.md`	Records native AgentX behavior.
`docs/reference/environment-variables.md`	Documents native controls and limits.
`docs/reference/authentication.md`	Documents dependency path ownership.
`docs/how-to/optimize.md`	Adds native AgentX instructions.
`docs/conceptual/optimization-loop.md`	Documents measurement-only phase behavior.
`docs/components/magpie.md`	Updates Magpie integration details.
`docs/compatibility.rst`	Records validated dependency pins.

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

chaojhou

Blocker: several native-only refusals are gated on AgentX in general, so they also switch off optimization for legacy AgentX sessions.

Native mode is opt-in (benchmark.agentx: enable), and SKILL.md keeps HYPERLOOM_AGENTX as the supported legacy switch. Most of the PR scopes its limits correctly with native_agentx_enabled(bench["agentx"]) (preflight, grid runner, benchmark runtime, CLI). These do not:

kernel.py:1050 is if agentx:, where agentx = is_agentx_mode(benchmark_mode) if benchmark_mode else agentx_enabled() (:1011). The GEAK phase is skipped for every AgentX session, and the legacy AgentX GEAK branches (proposal proxies with canonical revalidation in Hyperloom) are deleted.
integrate_patch.py:2350, request_handlers.py:4778, baseline.py:2366 / :2445 and writeback.py:6190 use agentx_active(...), which is HYPERLOOM_AGENTX or benchmark_mode == "agentx", unchanged from main.

Their messages all say "Native AgentX cannot ...", but a legacy HYPERLOOM_AGENTX=1 session -- today's production AgentX path -- would lose GEAK, integrate_patch source/kernel mutations, kernel mutation benchmarks and runtime_override, leaving measurement plus explore. The description does not mention this.

Requested change: gate these on native mode (the same native_agentx_enabled check on the session's source/accepted config), or, if legacy AgentX is meant to lose these levers too, say so in the description and get that signed off.

Must-fix: no GPU run. This bumps Magpie for every session, including synthetic ones, drops the install-time compat patch step (the runtime equivalents in preflight and baseline remain), and adds the native path, but per the description no GPU workload was executed, and ci-e2e has not run on this branch. Before merge, please run at least one synthetic GPU E2E and the shipped GLM-5.2 native example, and post the results.

haofrank · 2026-09-29T04:10:09Z

Review follow-up: legacy HYPERLOOM_AGENTX=1 optimization is preserved; the native-only refusals now follow explicit native session identity. Server tuning, GEAK handoff/revalidation, source/kernel integration, concurrency sweeps, and legacy resume retain their existing paths.

Both real BaselineExecutor GPU integration validations below passed on a newly allocated exclusive Crusoe 8×MI355X node. The full LLM optimizer E2E has not run: the operator's LLM credential configuration is unavailable. These baseline runs do not close that remaining E2E request, and no optimization gain is claimed.

Workload	Measured revisions	Result
Shipped GLM-5.2 native example, TP4/EP4, CONC8, HiCache + MTP	Hyperloom `2080dadd4`; Magpie release `5d6aaa29`	286.78509 output tokens/s across 4 GPUs, or 71.69627 tokens/s/GPU; P50 E2E-normalized interactivity 72.13054 tokens/s/user.
Synthetic Qwen3-0.6B BF16, TP1/EP1, CONC64, ISL/OSL 1024/1024	Current Hyperloom `0ef71137f`; Magpie hotfix `a3339dc`	320/320 throughput requests, 15,637.91443 output tokens/s on one GPU; real GSM8K strict-match 43/64 = 67.1875%.

The synthetic retry completed at 2026-09-29 04:03 UTC, with RUN_EVAL=true, Magpie exit code 0, and unchanged Torch/SGLang packages. Its 64 unique sample IDs and per-sample scores reproduce the reported accuracy. This is a 64-example smoke slice, not the full 1,319-example GSM8K evaluation. KV telemetry was unavailable (/metrics returned 404); throughput and accuracy artifacts were present.

The first synthetic attempt completed throughput but failed GSM8K because Magpie changed the working directory after sourcing InferenceX by a relative path. Fixed upstream in Magpie #113, currently draft and unmerged, and pinned the immutable fix in this PR. The v0.3.0 tag remains unchanged. Only ten generic script source blocks changed; native AgentX and shared runner Python code are byte-identical. The new source/wheel execution manifest was independently audited, and the old native manifest remains accepted for resume.

Native accounting caveats: the 3600-second sending window issued 990 requests; 989 completed, one tail request was cancelled, and there were zero recorded request errors. After the 30-second drain and a further 10-second cancellation-credit wait, AIPerf logged an ERROR and forced phase completion. This is not 990/990 completion. The 87 warmup records are excluded. All native schema/protocol, measurement, publishability, and submission-validity checks nevertheless passed. P50 TTFT/TPOT were 539.43/11.79 ms. accuracy=1.0 is the recorded-request error-rate gate, not model-answer accuracy.

The native table uses Magpie/InferenceX's 3606.24392-second completed-record span. AIPerf uses a 3640.00082-second phase window including drain, yielding 284.12549 output tokens/s for the same run. Server metrics emitted counter-reset warnings, so cache/MTP counter statistics are diagnostic. Both parsers report the nonfatal raw_inferencex_result_used provenance marker; neither successful validation relies on salvaging a nonzero process exit.

At current head 0ef71137f32e5775f133426d752f4e0555ec9c0f, the newly installed Magpie hotfix passed native identity/materialization checks, and the current parser reproduced the original native measurement exactly. That is a new preflight/reparse, not a repeated one-hour GPU measurement. Native logs/reports now display per-GPU values using the recorded physical GPU count, while keeping raw scoring values unchanged.

Reproduction and evidence:

Shipped config: examples/hyperloom-agentx-glm-5.2-mi355x/benchmark.yaml; mounted models: /shared_nfs/models/GLM-5.2-MXFP4 and /shared_nfs/models/Qwen3-0.6B.
Image: lmsysorg/sglang-rocm:v0.5.16-rocm720-mi35x-20260728, digest sha256:a95431225d5079bb398a3f49e7049630f62a2489197fbfd3c5ff84b02f026145.
InferenceX: 3d5581562f643f9bdeb8410cd924e2c70906c966; AIPerf: 754356e9a39acc6cc6afb242d123bb57c3fb6f75. Magpie release/hotfix: 5d6aaa29f2ee78e53ec35c8bc2e082d8a10c47f0 / a3339dc2776ee0c977fb3313fe89f56da7a91555.
NFS evidence: /shared_nfs/haoli103/hyperloom-pr1675-validation-20260929, sessions native/executor-session-02 and synthetic/executor-session-02; drivers in native/executor-validation.py and synthetic/executor-validation.py; current native check in native/hotfix-head-preflight-01 and native/hotfix-head-reparse.json.
Both task containers were removed and the new lease released after verifying no GPU processes remained. NFS evidence is retained.
Current-head full CI passed: Python 3.10/3.11 each total 24,391 passed / 81 skipped across six shards, with 91.27% coverage. Magpie hotfix CI also passed: 629 passed / 1 skipped on Python 3.10 and 3.12, with 90.16% coverage.

haofrank · 2026-10-06T03:57:14Z

Hyperloom-to-Magpie benchmark integration and repeated trace capture have been exercised on real GPUs.

Latest tested heads: Magpie cf80532 / Hyperloom 3b6c517 passed 266 / 295 focused tests. Magpie's Linux Python 3.10/3.12 CI passed. Hyperloom compatibility preparation leaves the audited Magpie package unchanged.
GPU integration: on GLM-5.2-MXFP4 / MI355X / SGLang (TP4, EP4, CONC8), baseline and candidate completed 1184/1184 and 591/591 measured requests with zero measured errors. The candidate was slower; no optimization gain is claimed.
Repeated profiling: completed 1202/1202 measured requests with zero measured errors and produced two captures, eight rank trace files.

GPU results above are from Magpie 938bfd7 / Hyperloom 80c589d, not the latest heads. Output-length warnings mean exact requested lengths remain unverified.

Profiling validates capture only: selected-rank quality checks reported blocked / warn; other ranks received file-inventory checks. Full TraceLens/Roofline analysis has not run, and profiled results cannot supply KEEP evidence. Latest-head native GPU and ordinary benchmark/GSM8K validation remain incomplete, as does the full Hyperloom optimization loop (Coordinator/KEEP/GEAK).

Adapt native AgentX promotion and GEAK revalidation to explicit orchestration collaborators, preserving launch identity, grading and rollback behavior. Update regression fixtures for main's state and grading contracts.

haofrank · 2026-10-06T05:09:16Z

Merged main (4ca5c7a) in e970c44; merge conflicts are resolved. Native AgentX promotion and GEAK revalidation now use main's current orchestration interfaces, preserving launch identity, grading and rollback behavior.

Validation: 3,048 regression tests passed, 10 skipped; repository-wide Ruff lint/format and shell syntax checks passed.

Installer tests did not pass in the local macOS environment; Linux CI is pending. GPU tests were not rerun. The PR remains Draft.

+            envs.pop(str(k), None)
+        for k, v in variant.extra_envs.items():
+            envs[str(k)] = str(v)
+        from ._workload_envs import pin_mlperf_round_concurrency


Ahmedhasssan-aig · 2026-10-07T19:27:28Z

Hi @haofrank

I was testing this PR using GLM 5.2 with AgentX to validate the end-to-end Hyperloom optimization flow. I have encountered an issue at the profiling stage. The profiling failed due to shape_discovery check. The check requires SGLang's profiler to support shape_discovery, including those inside CUDA graphs. SGLang 0.5.20 doesn't have that option, so Magpie refuses to profile at all. I tried to fix it, but Hyperloom treats any edited tool as tampered, and it reinstalled the original version, and profiling failed again on the same check.

I have raised an issue ticket on the Magpie repo "AMD-AGI/Magpie#116"

haofrank · 2026-10-07T22:19:26Z

@Ahmedhasssan-aig Magpie #117 is merged. This PR now pins d80eb4d with its updated audited fingerprint in d9bb74c. Shape discovery is optional, and per-batch graph export is preserved.

315 regression tests and the installed-package health check passed. GPU revalidation remains pending.

Preserve native launch configuration while adopting framework registry and KEEP helper refactors. Retire tests for the removed CK blockscale patch.

haofrank · 2026-10-08T17:38:14Z

@chaojhou Could you also review this PR? It integrates Magpie AgentX into Hyperloom and now pins the merged Magpie #117 fix. Thanks!

chaojhou · 2026-10-09T00:04:57Z

Re-reviewed at d4ccb15.

Magpie #114 and #117 are merged, and the pin (d80eb4d3) is now on Magpie main, so my earlier dependency blocker is resolved. On the default switch: fresh HYPERLOOM_AGENTX=1 launches selecting native AgentX is accepted as the product direction, with the condition below.

Two must-fix items, plus that condition:

1. A successful native profile, run as a standalone profile task, is recorded as a failure, and its trace is dropped.

_managed_profile_result.py:71-79 always returns valid_measurement=False (diagnostic only). That holds even when the traces were captured and status is succeeded.
writeback.py:1288-1292 promotes a profile result only when is_valid_measurement() passes. So the result goes to handle_unpromotable_result: the attempt is logged as failed, and the trace never reaches trace analysis.
Where this hits:
- a PRELUDE analysis that runs as a profile task (roofline disabled);
- any LLM-proposed profile action.
Each retry costs another full native run. The roofline path is not affected, because it reads main_trace_path directly.
Suggested fix: in the profile promotion path, accept a diagnostic profile result with trace_input_ready=True instead of requiring a valid measurement.

2. install.sh applies the native-AgentX Magpie audit to every install, including synthetic-only ones.

ensure_magpie (install.sh:1521-1543) requires AgentXConfig, the audited execution tree and an exact commit match. Under set -euo pipefail, a failed probe after the forced reinstall aborts the install. That contradicts install.sh:188-190 ("Generic benchmarks keep the importability contract").
These synthetic-only setups can no longer install, where import Magpie used to be enough:
- MAGPIE_REF pinned to a ref without AgentX support;
- a fork or a local MAGPIE_PACKAGE_SPEC;
- a preinstalled Magpie on an offline node.
The default online path passes.
Suggested fix: run the native capability probe only for native AgentX (or make it a warning for generic installs), and keep import Magpie as the generic contract.

Condition for the default switch: legacy AgentX settings that the native path does not honour should be refused at CLI preflight with a clear message, and documented. Today:

AIPERF_*, AGENTIC_* and WEKA_LOADER_OVERRIDE are scrubbed from the run environment without a word (agentx/native.py:655).
AGENTX_DATASET and AGENTX_WARMUP_REQUESTS_PER_LANE are rejected only when the first baseline is materialized (_workload_envs.py:745-758), which surfaces as a baseline failure.

CI: the one failing test (kernelforge/tests/roofline_ceiling/test_analyst.py) comes from #1603 via the main merge. This PR doesn't touch kernelforge, and main passes at the same base, so I've re-run the failed jobs. I've also added run-e2e so that CI E2E runs on this head.

github-actions · 2026-10-09T04:18:19Z

CI E2E report — ✅ Succeeded

item	value
result	✅ Succeeded
model	`Qwen/Qwen3-0.6B` (dense)
resources	1× GPU, TP=1
PR branch	`feature/magpie-0.3.0-agentx`
commit	`dd02d4ea303ac9e1923afd49d991e1f9505bf1ca`
session_id	`hyperloom:oneoff-91d337455ec24993`
backend run	`5dd959a7-4ab9-41d9-839a-10cb8a80c17d` on `-`
queue → dispatch	0s
run time	125m 55s
total	125m 55s

details

haofrank · 2026-10-09T20:27:51Z

Thanks @chaojhou, fixed all three in dd02d4e: successful diagnostic profiles now reach trace analysis without becoming KEEP evidence; the installer shares the CLI's mode-specific Magpie check; and unsupported legacy replay settings fail early with a clear preflight error.

Unit tests and coverage and the single-GPU Qwen3-0.6B smoke E2E are green on this head. The full native AgentX optimization E2E is still pending. Ready for another look when you have time.

haofrank added 2 commits September 24, 2026 22:36

feat(inference): use native Magpie AgentX recipes

2e6d8f4

test: align AgentX installer and kernel grading contracts

b51b4ae

github-code-quality Bot found potential problems Sep 25, 2026

View reviewed changes

Comment thread src/hyperloom/inference_optimizer/cli/__init__.py Fixed

github-advanced-security Bot found potential problems Sep 25, 2026

View reviewed changes

Comment thread src/hyperloom/inference_optimizer/cli/__init__.py Fixed

haofrank added 5 commits September 24, 2026 22:52

test: exercise native AgentX CLI and profiling boundaries

7fa6314

docs: clarify direct Magpie fast-mode invocation

de239a9

fix: preserve benchmark configuration decode errors

b5c8b81

fix: pin Magpie AgentX integration to v0.3.0 release

947ef28

Merge main into Magpie native AgentX integration

dea3ff0

haofrank marked this pull request as ready for review September 29, 2026 00:06

haofrank requested review from a team, Ahmedhasssan-aig, chao-xu-spec, devalshahamd, iraj465, tsrikris and yueliu14 as code owners September 29, 2026 00:06

haofrank requested a balanced review from Copilot September 29, 2026 00:06

Copilot started reviewing on behalf of haofrank September 29, 2026 00:08 View session

Copilot AI reviewed Sep 29, 2026

View reviewed changes

Comment thread src/hyperloom/orchestrator/actions/executors/baseline.py Outdated

Comment thread src/hyperloom/orchestrator/phases/kernel.py Outdated

fix: address native AgentX review feedback

571bedb

chaojhou requested changes Sep 29, 2026

View reviewed changes

haofrank added 4 commits September 28, 2026 18:45

Preserve legacy AgentX optimization with explicit native opt-in

2080dad

Fix AgentX mode dependency boundary and parallel test isolation

cfe108f

Display native throughput using recorded GPU counts

6a20d33

Document native throughput display fields

fb5e54a

haofrank mentioned this pull request Sep 29, 2026

Fix generic eval paths after changing the working directory AMD-AGI/Magpie#113

Merged

Pin Magpie generic eval path fix and audit its execution tree

0ef7113

Use native Magpie AgentX for new optimization sessions

6031d3a

haofrank added 9 commits October 2, 2026 10:20

Merge main and integrate reviewed Magpie AgentX profiling

9923b85

Preserve synthetic patch contracts with the pinned InferenceX layout

05fd8a6

Fix native AgentX identity across benchmark rounds

36a5eaa

Bind managed profile certification to verified workload parameters

654221e

Pin Magpie generic evaluation compatibility fix

ed35897

Preserve native AgentX identity at the final launch boundary

993c026

Pin Magpie phase-correct AgentX request accounting

1671967

Align native AgentX validation with measured request accounting

80c589d

Pin audited Magpie client compatibility fixes

3b6c517

Merge main into Magpie AgentX integration

e970c44

Adapt native AgentX promotion and GEAK revalidation to explicit orchestration collaborators, preserving launch identity, grading and rollback behavior. Update regression fixtures for main's state and grading contracts.

github-advanced-security Bot found potential problems Oct 6, 2026

View reviewed changes

Comment thread src/hyperloom/orchestrator/actions/executors/_grid_runner.py

envs.pop(str(k), None)

for k, v in variant.extra_envs.items():

envs[str(k)] = str(v)

from ._workload_envs import pin_mlperf_round_concurrency

haofrank marked this pull request as ready for review October 6, 2026 15:20

Merge main and preserve launch environment redaction

a6a2bbc

haofrank changed the title ~~Use Magpie native AgentX for Hyperloom optimization~~ feat(agentx) Use Magpie native AgentX for Hyperloom optimization Oct 8, 2026

Merge main and preserve native AgentX integration

98c276e

haofrank requested a review from chaojhou October 8, 2026 00:28

haofrank added 3 commits October 7, 2026 17:41

fix(agentx): keep shared request accounting below executor layer

7c9a585

Pin merged Magpie SGLang profiling compatibility fix

d9bb74c

Merge main into Magpie AgentX integration

d4ccb15

Preserve native launch configuration while adopting framework registry and KEEP helper refactors. Retire tests for the removed CK blockscale patch.

chaojhou added the run-e2e Run the single-GPU Hyperloom e2e smoke on this PR (again on every new commit while set) label Oct 9, 2026

haofrank added 2 commits October 9, 2026 10:32

Merge main and adapt native AgentX integration to relocated owners

6b47509

Fix native AgentX profile promotion and compatibility guards

dd02d4e

Conversation

haofrank commented Sep 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

chaojhou left a comment

Choose a reason for hiding this comment

Uh oh!

haofrank commented Sep 29, 2026

Uh oh!

haofrank commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

haofrank commented Oct 6, 2026

Uh oh!

Ahmedhasssan-aig commented Oct 7, 2026

Uh oh!

haofrank commented Oct 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

haofrank commented Oct 8, 2026

Uh oh!

chaojhou commented Oct 9, 2026

Uh oh!

github-actions Bot commented Oct 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CI E2E report — ✅ Succeeded

Uh oh!

haofrank commented Oct 9, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

haofrank commented Sep 25, 2026 •

edited

Loading

haofrank commented Oct 6, 2026 •

edited

Loading

haofrank commented Oct 7, 2026 •

edited

Loading

github-actions Bot commented Oct 9, 2026 •

edited

Loading