[CI/Build][ROCm] Move Pi0.5 CPU coverage to AMD nightly (#8559)

Signed-off-by: haic0 <149741444+haic0@users.noreply.github.com>
Signed-off-by: andyluo7 <andy.luo@amd.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: andyluo7 <andy.luo@amd.com>

.buildkite/amd/test-amd-merge.yml

@@ -50,18 +50,9 @@ steps:
         # that is not supported on CPU
         # Keep multi-GPU tests out of this one-GPU pool even when their module
         # also carries the cpu marker for unit tests.
+        # Keep π0.5 CPU units out of these shards: CUDA READY/MERGE already run
+        # the same suite mandatorily, and AMD covers it non-blocking in nightly.
         - "pytest -sv tests/diffusion -m 'core_model and cpu and not (cards_2 or cards_3 or cards_4 or cards_5 or cards_6 or cards_7 or cards_8)' --ignore=tests/diffusion/cache/test_teacache_extractors.py --ignore=tests/diffusion/models/pi05/test_pi05_units.py --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB"
-    # This module takes about 22 minutes on MI300 and was consistently assigned to shard 4,
-    # pushing that shard to the 45-minute timeout. Keep its module-scoped fixture in one lane.
-    - label: "Simple · Pi0.5 CPU Test"
-      agent_pool: mi300_1
-      depends_on: amd-build
-      mirror_hardwares: [amdproduction]
-      grade: Blocking
-      timeout_in_minutes: 45
-      commands:
-        - export VLLM_ROCM_USE_AITER=0
-        - "pytest -sv tests/diffusion/models/pi05/test_pi05_units.py -m 'core_model and cpu'"
 
     - label: "Diffusion · AuK VAE GPU Test"
       agent_pool: mi300_1

.buildkite/amd/test-amd-nightly.yml

@@ -83,6 +83,20 @@ steps:
         commands:
           - pytest -s -v tests/e2e/online_serving/test_cosyvoice3_tts_expansion.py -m "slow" --run-level "core_model"
 
+  # π0.5 CPU units are mandatory on CUDA READY/MERGE. On AMD they exercise the
+  # ROCm container/import path with CPU tensors only, so keep them off the
+  # blocking READY/MERGE critical path and retain non-blocking nightly coverage.
+  - group: ":robot_face: Pi0.5 CPU Nightly"
+    steps:
+      - label: "Simple · Pi0.5 CPU Test"
+        agent_pool: mi300_1
+        mirror_hardwares: [amdproduction]
+        grade: NonBlocking
+        timeout_in_minutes: 60
+        commands:
+          - export VLLM_ROCM_USE_AITER=0
+          - "pytest -sv tests/diffusion/models/pi05/test_pi05_units.py -m 'core_model and cpu'"
+
   - group: ":full_moon: LingBot-Video MI300 Nightly"
     steps:
       - label: "LingBot-Video Dense CFG-off Smoke"

.buildkite/amd/test-amd-ready.yml

@@ -50,18 +50,9 @@ steps:
         # that is not supported on CPU
         # Keep multi-GPU tests out of this one-GPU pool even when their module
         # also carries the cpu marker for unit tests.
+        # Keep π0.5 CPU units out of these shards: CUDA READY/MERGE already run
+        # the same suite mandatorily, and AMD covers it non-blocking in nightly.
         - "pytest -sv tests/diffusion -m 'core_model and cpu and not (cards_2 or cards_3 or cards_4 or cards_5 or cards_6 or cards_7 or cards_8)' --ignore=tests/diffusion/cache/test_teacache_extractors.py --ignore=tests/diffusion/models/pi05/test_pi05_units.py --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB"
-    # This module takes about 22 minutes on MI300 and was consistently assigned to shard 4,
-    # pushing that shard to the 45-minute timeout. Keep its module-scoped fixture in one lane.
-    - label: "Simple · Pi0.5 CPU Test"
-      agent_pool: mi300_1
-      depends_on: amd-build
-      mirror_hardwares: [amdproduction]
-      grade: Blocking
-      timeout_in_minutes: 45
-      commands:
-        - export VLLM_ROCM_USE_AITER=0
-        - "pytest -sv tests/diffusion/models/pi05/test_pi05_units.py -m 'core_model and cpu'"
 
     - label: "Diffusion · AuK VAE GPU Test"
       agent_pool: mi300_1

tests/buildkite/test_pi05_ci_routing.py

@@ -0,0 +1,81 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
+
+from pathlib import Path
+from shlex import split
+
+import pytest
+import yaml
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
+
+AMD_MERGE_PIPELINE = Path(".buildkite/amd/test-amd-merge.yml")
+AMD_NIGHTLY_PIPELINE = Path(".buildkite/amd/test-amd-nightly.yml")
+AMD_READY_PIPELINE = Path(".buildkite/amd/test-amd-ready.yml")
+CUDA_MERGE_PIPELINE = Path(".buildkite/cuda/test-merge.yml")
+CUDA_READY_PIPELINE = Path(".buildkite/cuda/test-ready.yml")
+PI05_LABEL = "Simple · Pi0.5 CPU Test"
+PI05_TEST_PATH = "tests/diffusion/models/pi05/test_pi05_units.py"
+
+
+def _pipeline_steps(pipeline_path: Path) -> list[dict]:
+    pipeline = yaml.safe_load(pipeline_path.read_text(encoding="utf-8"))
+    flattened = []
+
+    def walk(steps: list[dict]) -> None:
+        for step in steps:
+            flattened.append(step)
+            walk(step.get("steps", []))
+
+    walk(pipeline.get("steps", []))
+    return flattened
+
+
+def _find_step(label: str, pipeline_path: Path) -> dict:
+    step = next((step for step in _pipeline_steps(pipeline_path) if step.get("label") == label), None)
+    assert step is not None, f"missing pipeline step: {label}"
+    return step
+
+
+@pytest.mark.parametrize(
+    "pipeline_path",
+    [AMD_READY_PIPELINE, AMD_MERGE_PIPELINE],
+    ids=["ready", "merge"],
+)
+def test_pi05_cpu_suite_is_excluded_from_amd_pr_gates(pipeline_path: Path) -> None:
+    assert all(step.get("label") != PI05_LABEL for step in _pipeline_steps(pipeline_path))
+
+    step = _find_step("Simple · Diffusion Test · Shard %N/%t", pipeline_path)
+    pytest_command = next(command for command in step["commands"] if "pytest" in command)
+    assert f"--ignore={PI05_TEST_PATH}" in split(pytest_command)
+
+
+def test_pi05_cpu_suite_runs_nonblocking_in_amd_nightly() -> None:
+    steps = [step for step in _pipeline_steps(AMD_NIGHTLY_PIPELINE) if step.get("label") == PI05_LABEL]
+    assert len(steps) == 1
+
+    step = steps[0]
+    assert step["agent_pool"] == "mi300_1"
+    assert step["grade"] == "NonBlocking"
+    assert step["timeout_in_minutes"] == 60
+
+    pytest_command = next(command for command in step["commands"] if "pytest" in command)
+    argv = split(pytest_command)
+    assert PI05_TEST_PATH in argv
+    assert argv[argv.index("-m") + 1] == "core_model and cpu"
+
+
+@pytest.mark.parametrize(
+    "pipeline_path",
+    [CUDA_READY_PIPELINE, CUDA_MERGE_PIPELINE],
+    ids=["ready", "merge"],
+)
+def test_cuda_pr_gates_retain_pi05_cpu_coverage(pipeline_path: Path) -> None:
+    step = _find_step("Simple · Diffusion Test", pipeline_path)
+    pytest_command = next(command for command in step["commands"] if "pytest" in command)
+    argv = split(pytest_command)
+
+    assert "source_file_dependencies" not in step
+    assert "tests/diffusion" in argv
+    assert argv[argv.index("-m") + 1] == "core_model and cpu"
+    assert f"--ignore={PI05_TEST_PATH}" not in argv